id	sid	tid	token	lemma	pos
bracis-33576	1	1	erasmo	erasmo	NOUN
bracis-33576	1	2	:	:	PUNCT
bracis-33576	1	3	leveraging	leverage	VERB
bracis-33576	1	4	large	large	ADJ
bracis-33576	1	5	language	language	NOUN
bracis-33576	1	6	models	model	NOUN
bracis-33576	1	7	for	for	ADP
bracis-33576	1	8	 	 	SPACE
bracis-33576	1	9	enhanced	enhance	VERB
bracis-33576	1	10	clustering	clustering	ADJ
bracis-33576	1	11	segmentation	segmentation	NOUN
bracis-33576	1	12	|	|	ADV
bracis-33576	1	13	springer	springer	NOUN
bracis-33576	1	14	nature	nature	PROPN
bracis-33576	1	15	link	link	PROPN
bracis-33576	1	16	(	(	PUNCT
bracis-33576	1	17	formerly	formerly	ADV
bracis-33576	1	18	springerlink	springerlink	NOUN
bracis-33576	1	19	)	)	PUNCT
bracis-33576	1	20	skip	skip	VERB
bracis-33576	1	21	to	to	ADP
bracis-33576	1	22	main	main	ADJ
bracis-33576	1	23	content	content	NOUN
bracis-33576	1	24	advertisement	advertisement	NOUN
bracis-33576	1	25	log	log	NOUN
bracis-33576	1	26	in	in	ADP
bracis-33576	1	27	menu	menu	NOUN
bracis-33576	1	28	find	find	VERB
bracis-33576	1	29	a	a	DET
bracis-33576	1	30	journal	journal	NOUN
bracis-33576	1	31	publish	publish	VERB
bracis-33576	1	32	with	with	ADP
bracis-33576	1	33	us	we	PRON
bracis-33576	1	34	track	track	VERB
bracis-33576	1	35	your	your	PRON
bracis-33576	1	36	research	research	NOUN
bracis-33576	1	37	search	search	NOUN
bracis-33576	1	38	cart	cart	NOUN
bracis-33576	1	39	home	home	NOUN
bracis-33576	1	40	intelligent	intelligent	ADJ
bracis-33576	1	41	systems	system	NOUN
bracis-33576	1	42	conference	conference	NOUN
bracis-33576	1	43	paper	paper	NOUN
bracis-33576	1	44	erasmo	erasmo	PROPN
bracis-33576	1	45	:	:	PUNCT
bracis-33576	1	46	leveraging	leverage	VERB
bracis-33576	1	47	large	large	ADJ
bracis-33576	1	48	language	language	NOUN
bracis-33576	1	49	models	model	NOUN
bracis-33576	1	50	for	for	ADP
bracis-33576	1	51	 	 	SPACE
bracis-33576	1	52	enhanced	enhance	VERB
bracis-33576	1	53	clustering	clustering	ADJ
bracis-33576	1	54	segmentation	segmentation	NOUN
bracis-33576	1	55	conference	conference	NOUN
bracis-33576	1	56	paper	paper	NOUN
bracis-33576	1	57	first	first	ADV
bracis-33576	1	58	online	online	ADV
bracis-33576	1	59	:	:	PUNCT
bracis-33576	1	60	30	30	NUM
bracis-33576	1	61	january	january	NOUN
bracis-33576	1	62	2025	2025	NUM
bracis-33576	1	63	pp	pp	ADP
bracis-33576	1	64	414–429	414–429	NUM
bracis-33576	1	65	cite	cite	VERB
bracis-33576	1	66	this	this	DET
bracis-33576	1	67	conference	conference	NOUN
bracis-33576	1	68	paper	paper	NOUN
bracis-33576	1	69	access	access	NOUN
bracis-33576	1	70	provided	provide	VERB
bracis-33576	1	71	by	by	ADP
bracis-33576	1	72	university	university	PROPN
bracis-33576	1	73	of	of	ADP
bracis-33576	1	74	notre	notre	PROPN
bracis-33576	1	75	dame	dame	PROPN
bracis-33576	1	76	hesburgh	hesburgh	PROPN
bracis-33576	1	77	library	library	PROPN
bracis-33576	1	78	download	download	PROPN
bracis-33576	1	79	book	book	NOUN
bracis-33576	1	80	pdf	pdf	PROPN
bracis-33576	1	81	download	download	NOUN
bracis-33576	1	82	book	book	NOUN
bracis-33576	1	83	epub	epub	PROPN
bracis-33576	1	84	intelligent	intelligent	ADJ
bracis-33576	1	85	systems	system	NOUN
bracis-33576	1	86	(	(	PUNCT
bracis-33576	1	87	bracis	bracis	NOUN
bracis-33576	1	88	2024	2024	NUM
bracis-33576	1	89	)	)	PUNCT
bracis-33576	1	90	erasmo	erasmo	NOUN
bracis-33576	1	91	:	:	PUNCT
bracis-33576	1	92	leveraging	leverage	VERB
bracis-33576	1	93	large	large	ADJ
bracis-33576	1	94	language	language	NOUN
bracis-33576	1	95	models	model	NOUN
bracis-33576	1	96	for	for	ADP
bracis-33576	1	97	 	 	SPACE
bracis-33576	1	98	enhanced	enhance	VERB
bracis-33576	1	99	clustering	clustering	ADJ
bracis-33576	1	100	segmentation	segmentation	NOUN
bracis-33576	1	101	download	download	NOUN
bracis-33576	1	102	book	book	NOUN
bracis-33576	1	103	pdf	pdf	PROPN
bracis-33576	1	104	download	download	NOUN
bracis-33576	1	105	book	book	NOUN
bracis-33576	1	106	epub	epub	PROPN
bracis-33576	1	107	fillipe	fillipe	PROPN
bracis-33576	1	108	dos	dos	PROPN
bracis-33576	1	109	santos	santos	PROPN
bracis-33576	1	110	silva9,10,11	silva9,10,11	PROPN
bracis-33576	1	111	,	,	PUNCT
bracis-33576	1	112	gabriel	gabriel	PROPN
bracis-33576	1	113	kenzo	kenzo	PROPN
bracis-33576	1	114	kakimoto9,10,12	kakimoto9,10,12	PROPN
bracis-33576	1	115	,	,	PUNCT
bracis-33576	1	116	julio	julio	PROPN
bracis-33576	1	117	cesar	cesar	PROPN
bracis-33576	1	118	dos	do	VERB
bracis-33576	1	119	reis9,11	reis9,11	PROPN
bracis-33576	1	120	&	&	CCONJ
bracis-33576	1	121	…	…	PUNCT
bracis-33576	1	122	marcelo	marcelo	PROPN
bracis-33576	1	123	s.	s.	PROPN
bracis-33576	1	124	reis9,10,11	reis9,10,11	PROPN
bracis-33576	1	125	  	  	SPACE
bracis-33576	1	126	show	show	VERB
bracis-33576	1	127	authors	author	NOUN
bracis-33576	1	128	part	part	NOUN
bracis-33576	1	129	of	of	ADP
bracis-33576	1	130	the	the	DET
bracis-33576	1	131	book	book	NOUN
bracis-33576	1	132	series	series	NOUN
bracis-33576	1	133	:	:	PUNCT
bracis-33576	1	134	lecture	lecture	NOUN
bracis-33576	1	135	notes	note	NOUN
bracis-33576	1	136	in	in	ADP
bracis-33576	1	137	computer	computer	NOUN
bracis-33576	1	138	science	science	NOUN
bracis-33576	1	139	(	(	PUNCT
bracis-33576	1	140	(	(	PUNCT
bracis-33576	1	141	lnai	lnai	ADJ
bracis-33576	1	142	,	,	PUNCT
bracis-33576	1	143	volume	volume	NOUN
bracis-33576	1	144	15412	15412	NUM
bracis-33576	1	145	)	)	PUNCT
bracis-33576	1	146	)	)	PUNCT
bracis-33576	1	147	included	include	VERB
bracis-33576	1	148	in	in	ADP
bracis-33576	1	149	the	the	DET
bracis-33576	1	150	following	follow	VERB
bracis-33576	1	151	conference	conference	NOUN
bracis-33576	1	152	series	series	NOUN
bracis-33576	1	153	:	:	PUNCT
bracis-33576	1	154	brazilian	brazilian	ADJ
bracis-33576	1	155	conference	conference	NOUN
bracis-33576	1	156	on	on	ADP
bracis-33576	1	157	intelligent	intelligent	ADJ
bracis-33576	1	158	systems	system	NOUN
bracis-33576	1	159	512	512	NUM
bracis-33576	1	160	accesses	access	NOUN
bracis-33576	1	161	1	1	NUM
bracis-33576	1	162	citation	citation	NOUN
bracis-33576	1	163	abstract	abstract	ADJ
bracis-33576	1	164	cluster	cluster	NOUN
bracis-33576	1	165	analysis	analysis	NOUN
bracis-33576	1	166	plays	play	VERB
bracis-33576	1	167	a	a	DET
bracis-33576	1	168	crucial	crucial	ADJ
bracis-33576	1	169	role	role	NOUN
bracis-33576	1	170	in	in	ADP
bracis-33576	1	171	various	various	ADJ
bracis-33576	1	172	domains	domain	NOUN
bracis-33576	1	173	and	and	CCONJ
bracis-33576	1	174	applications	application	NOUN
bracis-33576	1	175	,	,	PUNCT
bracis-33576	1	176	such	such	ADJ
bracis-33576	1	177	as	as	ADP
bracis-33576	1	178	customer	customer	NOUN
bracis-33576	1	179	segmentation	segmentation	NOUN
bracis-33576	1	180	in	in	ADP
bracis-33576	1	181	marketing	marketing	NOUN
bracis-33576	1	182	.	.	PUNCT
bracis-33576	2	1	these	these	DET
bracis-33576	2	2	contexts	context	NOUN
bracis-33576	2	3	often	often	ADV
bracis-33576	2	4	involve	involve	VERB
bracis-33576	2	5	multimodal	multimodal	ADJ
bracis-33576	2	6	data	datum	NOUN
bracis-33576	2	7	,	,	PUNCT
bracis-33576	2	8	including	include	VERB
bracis-33576	2	9	both	both	CCONJ
bracis-33576	2	10	tabular	tabular	NOUN
bracis-33576	2	11	and	and	CCONJ
bracis-33576	2	12	textual	textual	ADJ
bracis-33576	2	13	datasets	dataset	NOUN
bracis-33576	2	14	,	,	PUNCT
bracis-33576	2	15	making	make	VERB
bracis-33576	2	16	it	it	PRON
bracis-33576	2	17	challenging	challenging	ADJ
bracis-33576	2	18	to	to	PART
bracis-33576	2	19	represent	represent	VERB
bracis-33576	2	20	hidden	hidden	ADJ
bracis-33576	2	21	patterns	pattern	NOUN
bracis-33576	2	22	for	for	ADP
bracis-33576	2	23	obtaining	obtain	VERB
bracis-33576	2	24	meaningful	meaningful	ADJ
bracis-33576	2	25	clusters	cluster	NOUN
bracis-33576	2	26	.	.	PUNCT
bracis-33576	3	1	this	this	DET
bracis-33576	3	2	study	study	NOUN
bracis-33576	3	3	introduces	introduce	VERB
bracis-33576	3	4	erasmo	erasmo	PROPN
bracis-33576	3	5	,	,	PUNCT
bracis-33576	3	6	a	a	DET
bracis-33576	3	7	framework	framework	NOUN
bracis-33576	3	8	designed	design	VERB
bracis-33576	3	9	to	to	ADP
bracis-33576	3	10	fine	fine	ADJ
bracis-33576	3	11	-	-	PUNCT
bracis-33576	3	12	tune	tune	NOUN
bracis-33576	3	13	a	a	DET
bracis-33576	3	14	pretrained	pretraine	VERB
bracis-33576	3	15	language	language	NOUN
bracis-33576	3	16	model	model	NOUN
bracis-33576	3	17	on	on	ADP
bracis-33576	3	18	textually	textually	ADV
bracis-33576	3	19	encoded	encode	VERB
bracis-33576	3	20	tabular	tabular	PROPN
bracis-33576	3	21	data	datum	NOUN
bracis-33576	3	22	and	and	CCONJ
bracis-33576	3	23	generate	generate	VERB
bracis-33576	3	24	embeddings	embedding	NOUN
bracis-33576	3	25	from	from	ADP
bracis-33576	3	26	the	the	DET
bracis-33576	3	27	fine	fine	ADV
bracis-33576	3	28	-	-	PUNCT
bracis-33576	3	29	tuned	tune	VERB
bracis-33576	3	30	model	model	NOUN
bracis-33576	3	31	.	.	PUNCT
bracis-33576	4	1	erasmo	erasmo	PROPN
bracis-33576	4	2	employs	employ	VERB
bracis-33576	4	3	a	a	DET
bracis-33576	4	4	textual	textual	ADJ
bracis-33576	4	5	converter	converter	NOUN
bracis-33576	4	6	to	to	PART
bracis-33576	4	7	transform	transform	VERB
bracis-33576	4	8	tabular	tabular	NOUN
bracis-33576	4	9	data	datum	NOUN
bracis-33576	4	10	into	into	ADP
bracis-33576	4	11	a	a	DET
bracis-33576	4	12	textual	textual	ADJ
bracis-33576	4	13	format	format	NOUN
bracis-33576	4	14	,	,	PUNCT
bracis-33576	4	15	enabling	enable	VERB
bracis-33576	4	16	the	the	DET
bracis-33576	4	17	language	language	NOUN
bracis-33576	4	18	model	model	NOUN
bracis-33576	4	19	to	to	PART
bracis-33576	4	20	process	process	VERB
bracis-33576	4	21	and	and	CCONJ
bracis-33576	4	22	understand	understand	VERB
bracis-33576	4	23	the	the	DET
bracis-33576	4	24	data	datum	NOUN
bracis-33576	4	25	more	more	ADV
bracis-33576	4	26	effectively	effectively	ADV
bracis-33576	4	27	.	.	PUNCT
bracis-33576	5	1	additionally	additionally	ADV
bracis-33576	5	2	,	,	PUNCT
bracis-33576	5	3	erasmo	erasmo	PROPN
bracis-33576	5	4	produces	produce	VERB
bracis-33576	5	5	contextually	contextually	ADV
bracis-33576	5	6	rich	rich	ADJ
bracis-33576	5	7	and	and	CCONJ
bracis-33576	5	8	structurally	structurally	ADV
bracis-33576	5	9	representative	representative	ADJ
bracis-33576	5	10	embeddings	embedding	NOUN
bracis-33576	5	11	through	through	ADP
bracis-33576	5	12	techniques	technique	NOUN
bracis-33576	5	13	such	such	ADJ
bracis-33576	5	14	as	as	ADP
bracis-33576	5	15	random	random	ADJ
bracis-33576	5	16	feature	feature	NOUN
bracis-33576	5	17	sequence	sequence	NOUN
bracis-33576	5	18	shuffling	shuffling	NOUN
bracis-33576	5	19	and	and	CCONJ
bracis-33576	5	20	number	number	NOUN
bracis-33576	5	21	verbalization	verbalization	NOUN
bracis-33576	5	22	.	.	PUNCT
bracis-33576	6	1	extensive	extensive	ADJ
bracis-33576	6	2	experimental	experimental	ADJ
bracis-33576	6	3	evaluations	evaluation	NOUN
bracis-33576	6	4	were	be	AUX
bracis-33576	6	5	conducted	conduct	VERB
bracis-33576	6	6	using	use	VERB
bracis-33576	6	7	multiple	multiple	ADJ
bracis-33576	6	8	datasets	dataset	NOUN
bracis-33576	6	9	and	and	CCONJ
bracis-33576	6	10	baseline	baseline	ADJ
bracis-33576	6	11	approaches	approach	NOUN
bracis-33576	6	12	.	.	PUNCT
bracis-33576	7	1	our	our	PRON
bracis-33576	7	2	results	result	NOUN
bracis-33576	7	3	demonstrate	demonstrate	VERB
bracis-33576	7	4	that	that	SCONJ
bracis-33576	7	5	erasmo	erasmo	PROPN
bracis-33576	7	6	fully	fully	ADV
bracis-33576	7	7	leverages	leverage	VERB
bracis-33576	7	8	the	the	DET
bracis-33576	7	9	specific	specific	ADJ
bracis-33576	7	10	context	context	NOUN
bracis-33576	7	11	of	of	ADP
bracis-33576	7	12	each	each	DET
bracis-33576	7	13	tabular	tabular	NOUN
bracis-33576	7	14	dataset	dataset	VERB
bracis-33576	7	15	,	,	PUNCT
bracis-33576	7	16	leading	lead	VERB
bracis-33576	7	17	to	to	ADP
bracis-33576	7	18	more	more	ADV
bracis-33576	7	19	precise	precise	ADJ
bracis-33576	7	20	and	and	CCONJ
bracis-33576	7	21	nuanced	nuanced	ADJ
bracis-33576	7	22	embeddings	embedding	NOUN
bracis-33576	7	23	for	for	ADP
bracis-33576	7	24	accurate	accurate	ADJ
bracis-33576	7	25	clustering	clustering	NOUN
bracis-33576	7	26	.	.	PUNCT
bracis-33576	8	1	this	this	DET
bracis-33576	8	2	approach	approach	NOUN
bracis-33576	8	3	enhances	enhance	VERB
bracis-33576	8	4	clustering	cluster	VERB
bracis-33576	8	5	performance	performance	NOUN
bracis-33576	8	6	by	by	ADP
bracis-33576	8	7	capturing	capture	VERB
bracis-33576	8	8	complex	complex	ADJ
bracis-33576	8	9	relationship	relationship	NOUN
bracis-33576	8	10	patterns	pattern	NOUN
bracis-33576	8	11	within	within	ADP
bracis-33576	8	12	diverse	diverse	ADJ
bracis-33576	8	13	tabular	tabular	PROPN
bracis-33576	8	14	data	datum	NOUN
bracis-33576	8	15	.	.	PUNCT
bracis-33576	9	1	access	access	NOUN
bracis-33576	9	2	provided	provide	VERB
bracis-33576	9	3	by	by	ADP
bracis-33576	9	4	university	university	PROPN
bracis-33576	9	5	of	of	ADP
bracis-33576	9	6	notre	notre	PROPN
bracis-33576	9	7	dame	dame	PROPN
bracis-33576	9	8	hesburgh	hesburgh	PROPN
bracis-33576	9	9	library	library	PROPN
bracis-33576	9	10	.	.	PUNCT
bracis-33576	10	1	download	download	PROPN
bracis-33576	10	2	conference	conference	NOUN
bracis-33576	10	3	paper	paper	NOUN
bracis-33576	10	4	pdf	pdf	NOUN
bracis-33576	10	5	similar	similar	ADJ
bracis-33576	10	6	content	content	NOUN
bracis-33576	10	7	being	be	AUX
bracis-33576	10	8	viewed	view	VERB
bracis-33576	10	9	by	by	ADP
bracis-33576	10	10	others	other	NOUN
bracis-33576	10	11	information	information	NOUN
bracis-33576	10	12	overload	overload	NOUN
bracis-33576	10	13	in	in	ADP
bracis-33576	10	14	crisis	crisis	NOUN
bracis-33576	10	15	management	management	NOUN
bracis-33576	10	16	:	:	PUNCT
bracis-33576	10	17	bilingual	bilingual	ADJ
bracis-33576	10	18	evaluation	evaluation	NOUN
bracis-33576	10	19	of	of	ADP
bracis-33576	10	20	embedding	embed	VERB
bracis-33576	10	21	models	model	NOUN
bracis-33576	10	22	for	for	ADP
bracis-33576	10	23	clustering	cluster	VERB
bracis-33576	10	24	social	social	ADJ
bracis-33576	10	25	media	medium	NOUN
bracis-33576	10	26	posts	post	NOUN
bracis-33576	10	27	in	in	ADP
bracis-33576	10	28	emergencies	emergency	NOUN
bracis-33576	10	29	chapter	chapter	NOUN
bracis-33576	10	30	©	©	PROPN
bracis-33576	10	31	2025	2025	NUM
bracis-33576	10	32	optimizing	optimize	VERB
bracis-33576	10	33	embedding	embed	VERB
bracis-33576	10	34	-	-	PUNCT
bracis-33576	10	35	based	base	VERB
bracis-33576	10	36	video	video	NOUN
bracis-33576	10	37	recommendation	recommendation	NOUN
bracis-33576	10	38	in	in	ADP
bracis-33576	10	39	spanish	spanish	ADJ
bracis-33576	10	40	using	use	VERB
bracis-33576	10	41	llms	llm	NOUN
bracis-33576	10	42	for	for	ADP
bracis-33576	10	43	automated	automate	VERB
bracis-33576	10	44	model	model	NOUN
bracis-33576	10	45	and	and	CCONJ
bracis-33576	10	46	parameter	parameter	NOUN
bracis-33576	10	47	selection	selection	NOUN
bracis-33576	10	48	chapter	chapter	NOUN
bracis-33576	10	49	©	©	PROPN
bracis-33576	10	50	2026	2026	NUM
bracis-33576	10	51	evaluating	evaluate	VERB
bracis-33576	10	52	the	the	DET
bracis-33576	10	53	 	 	SPACE
bracis-33576	10	54	impact	impact	NOUN
bracis-33576	10	55	of	of	ADP
bracis-33576	10	56	 	 	SPACE
bracis-33576	10	57	content	content	NOUN
bracis-33576	10	58	deletion	deletion	NOUN
bracis-33576	10	59	on	on	ADP
bracis-33576	10	60	 	 	SPACE
bracis-33576	10	61	tabular	tabular	PROPN
bracis-33576	10	62	data	datum	NOUN
bracis-33576	10	63	similarity	similarity	NOUN
bracis-33576	10	64	and	and	CCONJ
bracis-33576	10	65	 	 	SPACE
bracis-33576	10	66	retrieval	retrieval	NOUN
bracis-33576	10	67	using	use	VERB
bracis-33576	10	68	contextual	contextual	ADJ
bracis-33576	10	69	word	word	NOUN
bracis-33576	10	70	embeddings	embedding	NOUN
bracis-33576	10	71	chapter	chapter	NOUN
bracis-33576	10	72	©	©	NOUN
bracis-33576	10	73	2024	2024	NUM
bracis-33576	10	74	1	1	NUM
bracis-33576	10	75	introduction	introduction	NOUN
bracis-33576	10	76	tabular	tabular	NOUN
bracis-33576	10	77	data	datum	NOUN
bracis-33576	10	78	is	be	AUX
bracis-33576	10	79	ubiquitous	ubiquitous	ADJ
bracis-33576	10	80	in	in	ADP
bracis-33576	10	81	various	various	ADJ
bracis-33576	10	82	fields	field	NOUN
bracis-33576	10	83	such	such	ADJ
bracis-33576	10	84	as	as	ADP
bracis-33576	10	85	finance	finance	NOUN
bracis-33576	10	86	,	,	PUNCT
bracis-33576	10	87	healthcare	healthcare	PROPN
bracis-33576	10	88	,	,	PUNCT
bracis-33576	10	89	and	and	CCONJ
bracis-33576	10	90	marketing	marketing	NOUN
bracis-33576	10	91	,	,	PUNCT
bracis-33576	10	92	where	where	SCONJ
bracis-33576	10	93	it	it	PRON
bracis-33576	10	94	serves	serve	VERB
bracis-33576	10	95	as	as	ADP
bracis-33576	10	96	a	a	DET
bracis-33576	10	97	primary	primary	ADJ
bracis-33576	10	98	source	source	NOUN
bracis-33576	10	99	of	of	ADP
bracis-33576	10	100	information	information	NOUN
bracis-33576	10	101	for	for	ADP
bracis-33576	10	102	machine	machine	NOUN
bracis-33576	10	103	learning	learning	NOUN
bracis-33576	10	104	(	(	PUNCT
bracis-33576	10	105	ml	ml	NOUN
bracis-33576	10	106	)	)	PUNCT
bracis-33576	10	107	tasks	task	NOUN
bracis-33576	10	108	 	 	SPACE
bracis-33576	11	1	[	[	X
bracis-33576	11	2	15	15	NUM
bracis-33576	11	3	]	]	PUNCT
bracis-33576	11	4	.	.	PUNCT
bracis-33576	12	1	despite	despite	SCONJ
bracis-33576	12	2	its	its	PRON
bracis-33576	12	3	widespread	widespread	ADJ
bracis-33576	12	4	use	use	NOUN
bracis-33576	12	5	,	,	PUNCT
bracis-33576	12	6	extracting	extract	VERB
bracis-33576	12	7	meaningful	meaningful	ADJ
bracis-33576	12	8	insights	insight	NOUN
bracis-33576	12	9	from	from	ADP
bracis-33576	12	10	tabular	tabular	PROPN
bracis-33576	12	11	data	datum	NOUN
bracis-33576	12	12	remains	remain	VERB
bracis-33576	12	13	a	a	DET
bracis-33576	12	14	complex	complex	ADJ
bracis-33576	12	15	challenge	challenge	NOUN
bracis-33576	12	16	,	,	PUNCT
bracis-33576	12	17	particularly	particularly	ADV
bracis-33576	12	18	in	in	ADP
bracis-33576	12	19	clustering	clustering	ADJ
bracis-33576	12	20	tasks	task	NOUN
bracis-33576	12	21	 	 	SPACE
bracis-33576	13	1	[	[	X
bracis-33576	13	2	6	6	NUM
bracis-33576	13	3	]	]	PUNCT
bracis-33576	13	4	.	.	PUNCT
bracis-33576	14	1	these	these	DET
bracis-33576	14	2	challenges	challenge	NOUN
bracis-33576	14	3	include	include	VERB
bracis-33576	14	4	handling	handle	VERB
bracis-33576	14	5	heterogeneous	heterogeneous	ADJ
bracis-33576	14	6	feature	feature	NOUN
bracis-33576	14	7	types	type	NOUN
bracis-33576	14	8	,	,	PUNCT
bracis-33576	14	9	dealing	deal	VERB
bracis-33576	14	10	with	with	ADP
bracis-33576	14	11	high	high	ADJ
bracis-33576	14	12	-	-	PUNCT
bracis-33576	14	13	dimensional	dimensional	ADJ
bracis-33576	14	14	spaces	space	NOUN
bracis-33576	14	15	,	,	PUNCT
bracis-33576	14	16	and	and	CCONJ
bracis-33576	14	17	ensuring	ensure	VERB
bracis-33576	14	18	meaningful	meaningful	ADJ
bracis-33576	14	19	distance	distance	NOUN
bracis-33576	14	20	metrics	metric	NOUN
bracis-33576	14	21	.	.	PUNCT
bracis-33576	15	1	these	these	PRON
bracis-33576	15	2	can	can	AUX
bracis-33576	15	3	significantly	significantly	ADV
bracis-33576	15	4	impact	impact	VERB
bracis-33576	15	5	the	the	DET
bracis-33576	15	6	effectiveness	effectiveness	NOUN
bracis-33576	15	7	of	of	ADP
bracis-33576	15	8	clustering	clustering	ADJ
bracis-33576	15	9	algorithms	algorithm	NOUN
bracis-33576	15	10	in	in	ADP
bracis-33576	15	11	identifying	identify	VERB
bracis-33576	15	12	natural	natural	ADJ
bracis-33576	15	13	groupings	grouping	NOUN
bracis-33576	15	14	within	within	ADP
bracis-33576	15	15	the	the	DET
bracis-33576	15	16	data	datum	NOUN
bracis-33576	15	17	 	 	SPACE
bracis-33576	15	18	[	[	X
bracis-33576	15	19	14	14	NUM
bracis-33576	15	20	]	]	PUNCT
bracis-33576	15	21	.	.	PUNCT
bracis-33576	16	1	recent	recent	ADJ
bracis-33576	16	2	researchers	researcher	NOUN
bracis-33576	16	3	have	have	AUX
bracis-33576	16	4	explored	explore	VERB
bracis-33576	16	5	traditional	traditional	ADJ
bracis-33576	16	6	statistical	statistical	ADJ
bracis-33576	16	7	methods	method	NOUN
bracis-33576	16	8	and	and	CCONJ
bracis-33576	16	9	modern	modern	ADJ
bracis-33576	16	10	deep	deep	ADJ
bracis-33576	16	11	learning	learning	NOUN
bracis-33576	16	12	approaches	approach	NOUN
bracis-33576	16	13	 	 	SPACE
bracis-33576	17	1	[	[	X
bracis-33576	17	2	14	14	NUM
bracis-33576	17	3	,	,	PUNCT
bracis-33576	17	4	19	19	NUM
bracis-33576	17	5	,	,	PUNCT
bracis-33576	17	6	21	21	NUM
bracis-33576	17	7	]	]	PUNCT
bracis-33576	17	8	to	to	PART
bracis-33576	17	9	address	address	VERB
bracis-33576	17	10	these	these	DET
bracis-33576	17	11	challenges	challenge	NOUN
bracis-33576	17	12	.	.	PUNCT
bracis-33576	18	1	existing	exist	VERB
bracis-33576	18	2	studies	study	NOUN
bracis-33576	18	3	have	have	AUX
bracis-33576	18	4	utilized	utilize	VERB
bracis-33576	18	5	large	large	ADJ
bracis-33576	18	6	language	language	NOUN
bracis-33576	18	7	models	model	NOUN
bracis-33576	18	8	(	(	PUNCT
bracis-33576	18	9	llms	llm	NOUN
bracis-33576	18	10	)	)	PUNCT
bracis-33576	18	11	like	like	ADP
bracis-33576	18	12	openai	openai	PROPN
bracis-33576	18	13	’s	’s	PART
bracis-33576	18	14	gpt	gpt	NOUN
bracis-33576	18	15	and	and	CCONJ
bracis-33576	18	16	llama	llama	NOUN
bracis-33576	18	17	to	to	PART
bracis-33576	18	18	create	create	VERB
bracis-33576	18	19	embeddings	embedding	NOUN
bracis-33576	18	20	from	from	ADP
bracis-33576	18	21	textual	textual	ADJ
bracis-33576	18	22	datasets	dataset	NOUN
bracis-33576	18	23	,	,	PUNCT
bracis-33576	18	24	enhancing	enhance	VERB
bracis-33576	18	25	data	datum	NOUN
bracis-33576	18	26	representation	representation	NOUN
bracis-33576	18	27	and	and	CCONJ
bracis-33576	18	28	analysis	analysis	NOUN
bracis-33576	18	29	 	 	SPACE
bracis-33576	19	1	[	[	X
bracis-33576	19	2	6	6	NUM
bracis-33576	19	3	,	,	PUNCT
bracis-33576	19	4	14	14	NUM
bracis-33576	19	5	]	]	PUNCT
bracis-33576	19	6	.	.	PUNCT
bracis-33576	20	1	in	in	ADP
bracis-33576	20	2	addition	addition	NOUN
bracis-33576	20	3	,	,	PUNCT
bracis-33576	20	4	a	a	DET
bracis-33576	20	5	method	method	NOUN
bracis-33576	20	6	combining	combine	VERB
bracis-33576	20	7	llms	llm	NOUN
bracis-33576	20	8	and	and	CCONJ
bracis-33576	20	9	deterministic	deterministic	ADJ
bracis-33576	20	10	,	,	PUNCT
bracis-33576	20	11	independent	independent	ADJ
bracis-33576	20	12	-	-	PUNCT
bracis-33576	20	13	of	of	ADP
bracis-33576	20	14	-	-	PUNCT
bracis-33576	20	15	corpus	corpus	NOUN
bracis-33576	20	16	embeddings	embedding	NOUN
bracis-33576	20	17	(	(	PUNCT
bracis-33576	20	18	dice	dice	NOUN
bracis-33576	20	19	)	)	PUNCT
bracis-33576	20	20	has	have	AUX
bracis-33576	20	21	been	be	AUX
bracis-33576	20	22	proposed	propose	VERB
bracis-33576	20	23	to	to	PART
bracis-33576	20	24	generate	generate	VERB
bracis-33576	20	25	consistent	consistent	ADJ
bracis-33576	20	26	embeddings	embedding	NOUN
bracis-33576	20	27	across	across	ADP
bracis-33576	20	28	datasets	dataset	NOUN
bracis-33576	20	29	,	,	PUNCT
bracis-33576	20	30	improving	improve	VERB
bracis-33576	20	31	segmentation	segmentation	NOUN
bracis-33576	20	32	accuracy	accuracy	NOUN
bracis-33576	20	33	 	 	SPACE
bracis-33576	21	1	[	[	X
bracis-33576	21	2	19	19	NUM
bracis-33576	21	3	]	]	PUNCT
bracis-33576	21	4	.	.	PUNCT
bracis-33576	22	1	however	however	ADV
bracis-33576	22	2	,	,	PUNCT
bracis-33576	22	3	these	these	DET
bracis-33576	22	4	approaches	approach	NOUN
bracis-33576	22	5	often	often	ADV
bracis-33576	22	6	fail	fail	VERB
bracis-33576	22	7	to	to	PART
bracis-33576	22	8	fully	fully	ADV
bracis-33576	22	9	leverage	leverage	VERB
bracis-33576	22	10	the	the	DET
bracis-33576	22	11	specific	specific	ADJ
bracis-33576	22	12	context	context	NOUN
bracis-33576	22	13	of	of	ADP
bracis-33576	22	14	each	each	DET
bracis-33576	22	15	tabular	tabular	NOUN
bracis-33576	22	16	dataset	dataset	VERB
bracis-33576	22	17	,	,	PUNCT
bracis-33576	22	18	resulting	result	VERB
bracis-33576	22	19	in	in	ADP
bracis-33576	22	20	less	less	ADV
bracis-33576	22	21	nuanced	nuanced	ADJ
bracis-33576	22	22	embeddings	embedding	NOUN
bracis-33576	22	23	for	for	ADP
bracis-33576	22	24	precise	precise	ADJ
bracis-33576	22	25	clustering	clustering	NOUN
bracis-33576	22	26	.	.	PUNCT
bracis-33576	23	1	this	this	DET
bracis-33576	23	2	study	study	NOUN
bracis-33576	23	3	originally	originally	ADV
bracis-33576	23	4	introduces	introduce	VERB
bracis-33576	23	5	erasmo	erasmo	PROPN
bracis-33576	23	6	,	,	PUNCT
bracis-33576	23	7	our	our	PRON
bracis-33576	23	8	proposed	propose	VERB
bracis-33576	23	9	framework	framework	NOUN
bracis-33576	23	10	designed	design	VERB
bracis-33576	23	11	to	to	PART
bracis-33576	23	12	generate	generate	VERB
bracis-33576	23	13	high	high	ADJ
bracis-33576	23	14	-	-	PUNCT
bracis-33576	23	15	quality	quality	NOUN
bracis-33576	23	16	embeddings	embedding	NOUN
bracis-33576	23	17	from	from	ADP
bracis-33576	23	18	tabular	tabular	NOUN
bracis-33576	23	19	data	datum	NOUN
bracis-33576	23	20	using	use	VERB
bracis-33576	23	21	transformer	transformer	NOUN
bracis-33576	23	22	-	-	PUNCT
bracis-33576	23	23	based	base	VERB
bracis-33576	23	24	language	language	NOUN
bracis-33576	23	25	models	model	NOUN
bracis-33576	23	26	.	.	PUNCT
bracis-33576	24	1	these	these	DET
bracis-33576	24	2	embeddings	embedding	NOUN
bracis-33576	24	3	excel	excel	VERB
bracis-33576	24	4	in	in	ADP
bracis-33576	24	5	clustering	clustering	ADJ
bracis-33576	24	6	analysis	analysis	NOUN
bracis-33576	24	7	,	,	PUNCT
bracis-33576	24	8	revealing	reveal	VERB
bracis-33576	24	9	hidden	hide	VERB
bracis-33576	24	10	patterns	pattern	NOUN
bracis-33576	24	11	and	and	CCONJ
bracis-33576	24	12	groupings	grouping	NOUN
bracis-33576	24	13	.	.	PUNCT
bracis-33576	25	1	our	our	PRON
bracis-33576	25	2	solution	solution	NOUN
bracis-33576	25	3	can	can	AUX
bracis-33576	25	4	also	also	ADV
bracis-33576	25	5	be	be	AUX
bracis-33576	25	6	used	use	VERB
bracis-33576	25	7	in	in	ADP
bracis-33576	25	8	retrieval	retrieval	NOUN
bracis-33576	25	9	-	-	PUNCT
bracis-33576	25	10	augmented	augment	VERB
bracis-33576	25	11	generation	generation	NOUN
bracis-33576	25	12	(	(	PUNCT
bracis-33576	25	13	rag	rag	NOUN
bracis-33576	25	14	)	)	PUNCT
bracis-33576	25	15	systems	system	NOUN
bracis-33576	25	16	and	and	CCONJ
bracis-33576	25	17	other	other	ADJ
bracis-33576	25	18	tasks	task	NOUN
bracis-33576	25	19	to	to	PART
bracis-33576	25	20	gain	gain	VERB
bracis-33576	25	21	deeper	deep	ADJ
bracis-33576	25	22	insights	insight	NOUN
bracis-33576	25	23	from	from	ADP
bracis-33576	25	24	context	context	NOUN
bracis-33576	25	25	information	information	NOUN
bracis-33576	25	26	 	 	SPACE
bracis-33576	26	1	[	[	X
bracis-33576	26	2	7	7	NUM
bracis-33576	26	3	]	]	PUNCT
bracis-33576	26	4	.	.	PUNCT
bracis-33576	27	1	erasmo	erasmo	PROPN
bracis-33576	27	2	operates	operate	VERB
bracis-33576	27	3	through	through	ADP
bracis-33576	27	4	two	two	NUM
bracis-33576	27	5	stages	stage	NOUN
bracis-33576	27	6	:	:	PUNCT
bracis-33576	27	7	(	(	PUNCT
bracis-33576	27	8	1	1	X
bracis-33576	27	9	)	)	PUNCT
bracis-33576	27	10	fine	fine	ADV
bracis-33576	27	11	-	-	PUNCT
bracis-33576	27	12	tuning	tune	VERB
bracis-33576	27	13	a	a	DET
bracis-33576	27	14	pretrained	pretraine	VERB
bracis-33576	27	15	language	language	NOUN
bracis-33576	27	16	model	model	NOUN
bracis-33576	27	17	on	on	ADP
bracis-33576	27	18	textually	textually	ADV
bracis-33576	27	19	encoded	encode	VERB
bracis-33576	27	20	tabular	tabular	PROPN
bracis-33576	27	21	data	datum	NOUN
bracis-33576	27	22	;	;	PUNCT
bracis-33576	27	23	and	and	CCONJ
bracis-33576	27	24	(	(	PUNCT
bracis-33576	27	25	2	2	X
bracis-33576	27	26	)	)	PUNCT
bracis-33576	27	27	generating	generate	VERB
bracis-33576	27	28	embeddings	embedding	NOUN
bracis-33576	27	29	from	from	ADP
bracis-33576	27	30	the	the	DET
bracis-33576	27	31	fine	fine	ADV
bracis-33576	27	32	-	-	PUNCT
bracis-33576	27	33	tuned	tune	VERB
bracis-33576	27	34	model	model	NOUN
bracis-33576	27	35	.	.	PUNCT
bracis-33576	28	1	using	use	VERB
bracis-33576	28	2	techniques	technique	NOUN
bracis-33576	28	3	like	like	ADP
bracis-33576	28	4	random	random	ADJ
bracis-33576	28	5	feature	feature	NOUN
bracis-33576	28	6	sequence	sequence	NOUN
bracis-33576	28	7	shuffling	shuffling	NOUN
bracis-33576	28	8	and	and	CCONJ
bracis-33576	28	9	number	number	NOUN
bracis-33576	28	10	verbalization	verbalization	NOUN
bracis-33576	28	11	,	,	PUNCT
bracis-33576	28	12	erasmo	erasmo	PROPN
bracis-33576	28	13	produces	produce	VERB
bracis-33576	28	14	contextually	contextually	ADV
bracis-33576	28	15	rich	rich	ADJ
bracis-33576	28	16	and	and	CCONJ
bracis-33576	28	17	structurally	structurally	ADV
bracis-33576	28	18	representative	representative	ADJ
bracis-33576	28	19	embeddings	embedding	NOUN
bracis-33576	28	20	,	,	PUNCT
bracis-33576	28	21	outperforming	outperform	VERB
bracis-33576	28	22	all	all	DET
bracis-33576	28	23	clustering	clustering	ADJ
bracis-33576	28	24	strategies	strategy	NOUN
bracis-33576	28	25	from	from	ADP
bracis-33576	28	26	the	the	DET
bracis-33576	28	27	literature	literature	NOUN
bracis-33576	28	28	based	base	VERB
bracis-33576	28	29	on	on	ADP
bracis-33576	28	30	internal	internal	ADJ
bracis-33576	28	31	metrics	metric	NOUN
bracis-33576	28	32	.	.	PUNCT
bracis-33576	29	1	our	our	PRON
bracis-33576	29	2	experimental	experimental	ADJ
bracis-33576	29	3	evaluation	evaluation	NOUN
bracis-33576	29	4	used	use	VERB
bracis-33576	29	5	three	three	NUM
bracis-33576	29	6	clustering	cluster	VERB
bracis-33576	29	7	quality	quality	NOUN
bracis-33576	29	8	metrics	metric	NOUN
bracis-33576	29	9	to	to	PART
bracis-33576	29	10	compare	compare	VERB
bracis-33576	29	11	erasmo	erasmo	PROPN
bracis-33576	29	12	with	with	ADP
bracis-33576	29	13	state	state	NOUN
bracis-33576	29	14	-	-	PUNCT
bracis-33576	29	15	of	of	ADP
bracis-33576	29	16	-	-	PUNCT
bracis-33576	29	17	the	the	DET
bracis-33576	29	18	-	-	PUNCT
bracis-33576	29	19	art	art	NOUN
bracis-33576	29	20	methods	method	NOUN
bracis-33576	29	21	:	:	PUNCT
bracis-33576	29	22	silhouette	silhouette	NOUN
bracis-33576	29	23	score	score	NOUN
bracis-33576	29	24	(	(	PUNCT
bracis-33576	29	25	ss	ss	NOUN
bracis-33576	29	26	)	)	PUNCT
bracis-33576	29	27	,	,	PUNCT
bracis-33576	29	28	calinski	calinski	NOUN
bracis-33576	29	29	-	-	PUNCT
bracis-33576	29	30	harabasz	harabasz	NOUN
bracis-33576	29	31	index	index	NOUN
bracis-33576	29	32	(	(	PUNCT
bracis-33576	29	33	chi	chi	NOUN
bracis-33576	29	34	)	)	PUNCT
bracis-33576	29	35	,	,	PUNCT
bracis-33576	29	36	and	and	CCONJ
bracis-33576	29	37	davies	davy	NOUN
bracis-33576	29	38	-	-	PUNCT
bracis-33576	29	39	bouldin	bouldin	NOUN
bracis-33576	29	40	index	index	NOUN
bracis-33576	29	41	(	(	PUNCT
bracis-33576	29	42	dbi	dbi	PROPN
bracis-33576	29	43	)	)	PUNCT
bracis-33576	29	44	.	.	PUNCT
bracis-33576	30	1	these	these	DET
bracis-33576	30	2	metrics	metric	NOUN
bracis-33576	30	3	comprehensively	comprehensively	ADV
bracis-33576	30	4	assess	assess	VERB
bracis-33576	30	5	clustering	clustering	ADJ
bracis-33576	30	6	effectiveness	effectiveness	NOUN
bracis-33576	30	7	results	result	NOUN
bracis-33576	30	8	by	by	ADP
bracis-33576	30	9	measuring	measure	VERB
bracis-33576	30	10	cohesion	cohesion	NOUN
bracis-33576	30	11	,	,	PUNCT
bracis-33576	30	12	separation	separation	NOUN
bracis-33576	30	13	,	,	PUNCT
bracis-33576	30	14	and	and	CCONJ
bracis-33576	30	15	overall	overall	ADJ
bracis-33576	30	16	cluster	cluster	NOUN
bracis-33576	30	17	structure	structure	NOUN
bracis-33576	30	18	.	.	PUNCT
bracis-33576	31	1	we	we	PRON
bracis-33576	31	2	extensively	extensively	ADV
bracis-33576	31	3	evaluated	evaluate	VERB
bracis-33576	31	4	erasmo	erasmo	PROPN
bracis-33576	31	5	on	on	ADP
bracis-33576	31	6	real	real	ADJ
bracis-33576	31	7	-	-	PUNCT
bracis-33576	31	8	world	world	NOUN
bracis-33576	31	9	datasets	dataset	NOUN
bracis-33576	31	10	without	without	ADP
bracis-33576	31	11	true	true	ADJ
bracis-33576	31	12	labels	label	NOUN
bracis-33576	31	13	,	,	PUNCT
bracis-33576	31	14	including	include	VERB
bracis-33576	31	15	banking	banking	NOUN
bracis-33576	31	16	marketing	marketing	NOUN
bracis-33576	31	17	targets	target	NOUN
bracis-33576	31	18	,	,	PUNCT
bracis-33576	31	19	e	e	NOUN
bracis-33576	31	20	-	-	ADJ
bracis-33576	31	21	commerce	commerce	ADJ
bracis-33576	31	22	public	public	NOUN
bracis-33576	31	23	dataset	dataset	VERB
bracis-33576	31	24	by	by	ADP
bracis-33576	31	25	olist	olist	ADJ
bracis-33576	31	26	,	,	PUNCT
bracis-33576	31	27	yelp	yelp	NOUN
bracis-33576	31	28	reviews	review	NOUN
bracis-33576	31	29	,	,	PUNCT
bracis-33576	31	30	petfinder.my	petfinder.my	PROPN
bracis-33576	31	31	,	,	PUNCT
bracis-33576	31	32	and	and	CCONJ
bracis-33576	31	33	women	woman	NOUN
bracis-33576	31	34	’s	’s	PART
bracis-33576	31	35	clothing	clothing	NOUN
bracis-33576	31	36	reviews	review	NOUN
bracis-33576	31	37	.	.	PUNCT
bracis-33576	32	1	these	these	DET
bracis-33576	32	2	datasets	dataset	NOUN
bracis-33576	32	3	encompass	encompass	VERB
bracis-33576	32	4	diversified	diversified	ADJ
bracis-33576	32	5	information	information	NOUN
bracis-33576	32	6	and	and	CCONJ
bracis-33576	32	7	present	present	VERB
bracis-33576	32	8	various	various	ADJ
bracis-33576	32	9	challenges	challenge	NOUN
bracis-33576	32	10	,	,	PUNCT
bracis-33576	32	11	rigorously	rigorously	ADV
bracis-33576	32	12	testing	test	VERB
bracis-33576	32	13	erasmo	erasmo	PROPN
bracis-33576	32	14	’s	’s	PART
bracis-33576	32	15	clustering	clustering	ADJ
bracis-33576	32	16	capabilities	capability	NOUN
bracis-33576	32	17	.	.	PUNCT
bracis-33576	33	1	this	this	DET
bracis-33576	33	2	article	article	NOUN
bracis-33576	33	3	provides	provide	VERB
bracis-33576	33	4	the	the	DET
bracis-33576	33	5	main	main	ADJ
bracis-33576	33	6	contributions	contribution	NOUN
bracis-33576	33	7	as	as	SCONJ
bracis-33576	33	8	follows	follow	VERB
bracis-33576	33	9	:	:	PUNCT
bracis-33576	33	10	we	we	PRON
bracis-33576	33	11	introduce	introduce	VERB
bracis-33576	33	12	erasmo	erasmo	PROPN
bracis-33576	33	13	,	,	PUNCT
bracis-33576	33	14	a	a	DET
bracis-33576	33	15	novel	novel	ADJ
bracis-33576	33	16	framework	framework	NOUN
bracis-33576	33	17	that	that	PRON
bracis-33576	33	18	leverages	leverage	VERB
bracis-33576	33	19	transformer	transformer	NOUN
bracis-33576	33	20	-	-	PUNCT
bracis-33576	33	21	based	base	VERB
bracis-33576	33	22	language	language	NOUN
bracis-33576	33	23	models	model	NOUN
bracis-33576	33	24	to	to	PART
bracis-33576	33	25	generate	generate	VERB
bracis-33576	33	26	high	high	ADJ
bracis-33576	33	27	-	-	PUNCT
bracis-33576	33	28	quality	quality	NOUN
bracis-33576	33	29	embeddings	embedding	NOUN
bracis-33576	33	30	from	from	ADP
bracis-33576	33	31	tabular	tabular	PROPN
bracis-33576	33	32	data	datum	NOUN
bracis-33576	33	33	,	,	PUNCT
bracis-33576	33	34	enhancing	enhance	VERB
bracis-33576	33	35	clustering	clustering	ADJ
bracis-33576	33	36	analysis	analysis	NOUN
bracis-33576	33	37	.	.	PUNCT
bracis-33576	34	1	we	we	PRON
bracis-33576	34	2	demonstrate	demonstrate	VERB
bracis-33576	34	3	that	that	SCONJ
bracis-33576	34	4	our	our	PRON
bracis-33576	34	5	framework	framework	NOUN
bracis-33576	34	6	significantly	significantly	ADV
bracis-33576	34	7	improves	improve	VERB
bracis-33576	34	8	clustering	cluster	VERB
bracis-33576	34	9	performance	performance	NOUN
bracis-33576	34	10	by	by	ADP
bracis-33576	34	11	capturing	capture	VERB
bracis-33576	34	12	the	the	DET
bracis-33576	34	13	complex	complex	ADJ
bracis-33576	34	14	relationships	relationship	NOUN
bracis-33576	34	15	within	within	ADP
bracis-33576	34	16	tabular	tabular	PROPN
bracis-33576	34	17	data	datum	NOUN
bracis-33576	34	18	through	through	ADP
bracis-33576	34	19	random	random	ADJ
bracis-33576	34	20	feature	feature	NOUN
bracis-33576	34	21	sequence	sequence	NOUN
bracis-33576	34	22	shuffling	shuffling	NOUN
bracis-33576	34	23	and	and	CCONJ
bracis-33576	34	24	number	number	NOUN
bracis-33576	34	25	verbalization	verbalization	NOUN
bracis-33576	34	26	techniques	technique	NOUN
bracis-33576	34	27	.	.	PUNCT
bracis-33576	35	1	we	we	PRON
bracis-33576	35	2	experimentally	experimentally	ADV
bracis-33576	35	3	achieve	achieve	VERB
bracis-33576	35	4	state	state	NOUN
bracis-33576	35	5	-	-	PUNCT
bracis-33576	35	6	of	of	ADP
bracis-33576	35	7	-	-	PUNCT
bracis-33576	35	8	the	the	DET
bracis-33576	35	9	-	-	PUNCT
bracis-33576	35	10	art	art	NOUN
bracis-33576	35	11	clustering	clustering	NOUN
bracis-33576	35	12	results	result	NOUN
bracis-33576	35	13	with	with	ADP
bracis-33576	35	14	erasmo	erasmo	NOUN
bracis-33576	35	15	,	,	PUNCT
bracis-33576	35	16	showcasing	showcase	VERB
bracis-33576	35	17	its	its	PRON
bracis-33576	35	18	effectiveness	effectiveness	NOUN
bracis-33576	35	19	in	in	ADP
bracis-33576	35	20	identifying	identify	VERB
bracis-33576	35	21	patterns	pattern	NOUN
bracis-33576	35	22	and	and	CCONJ
bracis-33576	35	23	groupings	grouping	NOUN
bracis-33576	35	24	within	within	ADP
bracis-33576	35	25	diverse	diverse	ADJ
bracis-33576	35	26	tabular	tabular	PROPN
bracis-33576	35	27	datasets	dataset	NOUN
bracis-33576	35	28	.	.	PUNCT
bracis-33576	36	1	to	to	ADP
bracis-33576	36	2	the	the	DET
bracis-33576	36	3	best	good	ADJ
bracis-33576	36	4	of	of	ADP
bracis-33576	36	5	our	our	PRON
bracis-33576	36	6	knowledge	knowledge	NOUN
bracis-33576	36	7	,	,	PUNCT
bracis-33576	36	8	erasmo	erasmo	PROPN
bracis-33576	36	9	is	be	AUX
bracis-33576	36	10	the	the	DET
bracis-33576	36	11	first	first	ADJ
bracis-33576	36	12	framework	framework	NOUN
bracis-33576	36	13	to	to	PART
bracis-33576	36	14	fully	fully	ADV
bracis-33576	36	15	integrate	integrate	VERB
bracis-33576	36	16	and	and	CCONJ
bracis-33576	36	17	fine	fine	ADJ
bracis-33576	36	18	-	-	PUNCT
bracis-33576	36	19	tune	tune	NOUN
bracis-33576	36	20	transformer	transformer	NOUN
bracis-33576	36	21	-	-	PUNCT
bracis-33576	36	22	based	base	VERB
bracis-33576	36	23	language	language	NOUN
bracis-33576	36	24	models	model	NOUN
bracis-33576	36	25	specifically	specifically	ADV
bracis-33576	36	26	for	for	ADP
bracis-33576	36	27	generating	generate	VERB
bracis-33576	36	28	embeddings	embedding	NOUN
bracis-33576	36	29	from	from	ADP
bracis-33576	36	30	tabular	tabular	PROPN
bracis-33576	36	31	data	datum	NOUN
bracis-33576	36	32	,	,	PUNCT
bracis-33576	36	33	resulting	result	VERB
bracis-33576	36	34	in	in	ADP
bracis-33576	36	35	superior	superior	ADJ
bracis-33576	36	36	clustering	clustering	ADJ
bracis-33576	36	37	outcomes	outcome	NOUN
bracis-33576	36	38	.	.	PUNCT
bracis-33576	37	1	the	the	DET
bracis-33576	37	2	remainder	remainder	NOUN
bracis-33576	37	3	of	of	ADP
bracis-33576	37	4	this	this	DET
bracis-33576	37	5	article	article	NOUN
bracis-33576	37	6	is	be	AUX
bracis-33576	37	7	organized	organize	VERB
bracis-33576	37	8	as	as	SCONJ
bracis-33576	37	9	follows	follow	VERB
bracis-33576	37	10	.	.	PUNCT
bracis-33576	38	1	section	section	NOUN
bracis-33576	38	2	 	 	SPACE
bracis-33576	38	3	2	2	NUM
bracis-33576	38	4	presents	present	VERB
bracis-33576	38	5	a	a	DET
bracis-33576	38	6	synthesis	synthesis	NOUN
bracis-33576	38	7	of	of	ADP
bracis-33576	38	8	related	related	ADJ
bracis-33576	38	9	work	work	NOUN
bracis-33576	38	10	.	.	PUNCT
bracis-33576	39	1	section	section	NOUN
bracis-33576	39	2	 	 	SPACE
bracis-33576	39	3	3	3	NUM
bracis-33576	39	4	details	detail	NOUN
bracis-33576	39	5	the	the	DET
bracis-33576	39	6	erasmo	erasmo	PROPN
bracis-33576	39	7	framework	framework	NOUN
bracis-33576	39	8	.	.	PUNCT
bracis-33576	40	1	section	section	NOUN
bracis-33576	40	2	 	 	SPACE
bracis-33576	40	3	4	4	NUM
bracis-33576	40	4	outlines	outline	VERB
bracis-33576	40	5	our	our	PRON
bracis-33576	40	6	experimental	experimental	ADJ
bracis-33576	40	7	methodology	methodology	NOUN
bracis-33576	40	8	.	.	PUNCT
bracis-33576	41	1	section	section	NOUN
bracis-33576	41	2	 	 	SPACE
bracis-33576	41	3	5	5	NUM
bracis-33576	41	4	presents	present	VERB
bracis-33576	41	5	the	the	DET
bracis-33576	41	6	results	result	NOUN
bracis-33576	41	7	of	of	ADP
bracis-33576	41	8	our	our	PRON
bracis-33576	41	9	evaluations	evaluation	NOUN
bracis-33576	41	10	.	.	PUNCT
bracis-33576	42	1	section	section	NOUN
bracis-33576	42	2	 	 	SPACE
bracis-33576	42	3	6	6	NUM
bracis-33576	42	4	discusses	discuss	VERB
bracis-33576	42	5	the	the	DET
bracis-33576	42	6	implications	implication	NOUN
bracis-33576	42	7	of	of	ADP
bracis-33576	42	8	our	our	PRON
bracis-33576	42	9	findings	finding	NOUN
bracis-33576	42	10	.	.	PUNCT
bracis-33576	43	1	finally	finally	ADV
bracis-33576	43	2	,	,	PUNCT
bracis-33576	43	3	sect	sect	NOUN
bracis-33576	43	4	.	.	PUNCT
bracis-33576	43	5	 	 	SPACE
bracis-33576	43	6	7	7	NUM
bracis-33576	43	7	provides	provide	VERB
bracis-33576	43	8	conclusions	conclusion	NOUN
bracis-33576	43	9	and	and	CCONJ
bracis-33576	43	10	directions	direction	NOUN
bracis-33576	43	11	for	for	ADP
bracis-33576	43	12	future	future	ADJ
bracis-33576	43	13	work	work	NOUN
bracis-33576	43	14	.	.	PUNCT
bracis-33576	44	1	2	2	NUM
bracis-33576	44	2	related	relate	VERB
bracis-33576	44	3	work	work	NOUN
bracis-33576	44	4	several	several	ADJ
bracis-33576	44	5	studies	study	NOUN
bracis-33576	44	6	have	have	AUX
bracis-33576	44	7	explored	explore	VERB
bracis-33576	44	8	the	the	DET
bracis-33576	44	9	application	application	NOUN
bracis-33576	44	10	of	of	ADP
bracis-33576	44	11	llms	llm	NOUN
bracis-33576	44	12	to	to	PART
bracis-33576	44	13	transform	transform	VERB
bracis-33576	44	14	tabular	tabular	NOUN
bracis-33576	44	15	data	datum	NOUN
bracis-33576	44	16	for	for	ADP
bracis-33576	44	17	clustering	clustering	ADJ
bracis-33576	44	18	tasks	task	NOUN
bracis-33576	44	19	,	,	PUNCT
bracis-33576	44	20	demonstrating	demonstrate	VERB
bracis-33576	44	21	the	the	DET
bracis-33576	44	22	potential	potential	NOUN
bracis-33576	44	23	to	to	PART
bracis-33576	44	24	enhance	enhance	VERB
bracis-33576	44	25	user	user	NOUN
bracis-33576	44	26	segmentation	segmentation	NOUN
bracis-33576	44	27	and	and	CCONJ
bracis-33576	44	28	data	datum	NOUN
bracis-33576	44	29	analysis	analysis	NOUN
bracis-33576	44	30	 	 	SPACE
bracis-33576	45	1	[	[	X
bracis-33576	45	2	6	6	NUM
bracis-33576	45	3	,	,	PUNCT
bracis-33576	45	4	14	14	NUM
bracis-33576	45	5	,	,	PUNCT
bracis-33576	45	6	19	19	NUM
bracis-33576	45	7	,	,	PUNCT
bracis-33576	45	8	21	21	NUM
bracis-33576	45	9	,	,	PUNCT
bracis-33576	45	10	23	23	NUM
bracis-33576	45	11	,	,	PUNCT
bracis-33576	45	12	24	24	NUM
bracis-33576	45	13	]	]	PUNCT
bracis-33576	45	14	.	.	PUNCT
bracis-33576	46	1	zhu	zhu	PROPN
bracis-33576	46	2	et	et	PROPN
bracis-33576	46	3	al	al	PROPN
bracis-33576	46	4	.	.	PUNCT
bracis-33576	47	1	[	[	X
bracis-33576	47	2	24	24	NUM
bracis-33576	47	3	]	]	PUNCT
bracis-33576	47	4	proposed	propose	VERB
bracis-33576	47	5	a	a	DET
bracis-33576	47	6	novel	novel	ADJ
bracis-33576	47	7	method	method	NOUN
bracis-33576	47	8	named	name	VERB
bracis-33576	47	9	word	word	NOUN
bracis-33576	47	10	embedding	embed	VERB
bracis-33576	47	11	of	of	ADP
bracis-33576	47	12	dimensionality	dimensionality	NOUN
bracis-33576	47	13	reduction	reduction	NOUN
bracis-33576	47	14	(	(	PUNCT
bracis-33576	47	15	werd	werd	NOUN
bracis-33576	47	16	)	)	PUNCT
bracis-33576	47	17	for	for	ADP
bracis-33576	47	18	document	document	NOUN
bracis-33576	47	19	clustering	clustering	NOUN
bracis-33576	47	20	.	.	PUNCT
bracis-33576	48	1	their	their	PRON
bracis-33576	48	2	approach	approach	NOUN
bracis-33576	48	3	integrates	integrate	VERB
bracis-33576	48	4	pre	pre	ADJ
bracis-33576	48	5	-	-	ADJ
bracis-33576	48	6	trained	train	VERB
bracis-33576	48	7	word	word	NOUN
bracis-33576	48	8	embeddings	embedding	NOUN
bracis-33576	48	9	with	with	ADP
bracis-33576	48	10	dimensionality	dimensionality	NOUN
bracis-33576	48	11	reduction	reduction	NOUN
bracis-33576	48	12	techniques	technique	NOUN
bracis-33576	48	13	.	.	PUNCT
bracis-33576	49	1	in	in	ADP
bracis-33576	49	2	their	their	PRON
bracis-33576	49	3	work	work	NOUN
bracis-33576	49	4	,	,	PUNCT
bracis-33576	49	5	sentence	sentence	NOUN
bracis-33576	49	6	-	-	PUNCT
bracis-33576	49	7	bert	bert	PROPN
bracis-33576	49	8	embeds	embed	VERB
bracis-33576	49	9	them	they	PRON
bracis-33576	49	10	into	into	ADP
bracis-33576	49	11	high	high	ADJ
bracis-33576	49	12	-	-	PUNCT
bracis-33576	49	13	dimensional	dimensional	ADJ
bracis-33576	49	14	vectors	vector	NOUN
bracis-33576	49	15	after	after	ADP
bracis-33576	49	16	preprocessing	preprocesse	VERB
bracis-33576	49	17	documents	document	NOUN
bracis-33576	49	18	,	,	PUNCT
bracis-33576	49	19	which	which	DET
bracis-33576	49	20	pacmap	pacmap	NOUN
bracis-33576	49	21	then	then	ADV
bracis-33576	49	22	reduces	reduce	VERB
bracis-33576	49	23	.	.	PUNCT
bracis-33576	50	1	spectral	spectral	ADJ
bracis-33576	50	2	clustering	clustering	NOUN
bracis-33576	50	3	is	be	AUX
bracis-33576	50	4	applied	apply	VERB
bracis-33576	50	5	,	,	PUNCT
bracis-33576	50	6	followed	follow	VERB
bracis-33576	50	7	by	by	ADP
bracis-33576	50	8	non	non	ADJ
bracis-33576	50	9	-	-	ADJ
bracis-33576	50	10	negative	negative	ADJ
bracis-33576	50	11	matrix	matrix	NOUN
bracis-33576	50	12	factorization	factorization	NOUN
bracis-33576	50	13	to	to	PART
bracis-33576	50	14	extract	extract	VERB
bracis-33576	50	15	keywords	keyword	NOUN
bracis-33576	50	16	.	.	PUNCT
bracis-33576	51	1	clusterllm	clusterllm	VERB
bracis-33576	51	2	 	 	SPACE
bracis-33576	52	1	[	[	X
bracis-33576	52	2	23	23	NUM
bracis-33576	52	3	]	]	PUNCT
bracis-33576	52	4	,	,	PUNCT
bracis-33576	52	5	a	a	DET
bracis-33576	52	6	novel	novel	ADJ
bracis-33576	52	7	text	text	NOUN
bracis-33576	52	8	clustering	clustering	NOUN
bracis-33576	52	9	framework	framework	NOUN
bracis-33576	52	10	,	,	PUNCT
bracis-33576	52	11	leverages	leverage	VERB
bracis-33576	52	12	feedback	feedback	NOUN
bracis-33576	52	13	from	from	ADP
bracis-33576	52	14	llms	llm	NOUN
bracis-33576	52	15	such	such	ADJ
bracis-33576	52	16	as	as	ADP
bracis-33576	52	17	chatgpt	chatgpt	NOUN
bracis-33576	52	18	.	.	PUNCT
bracis-33576	53	1	this	this	DET
bracis-33576	53	2	method	method	NOUN
bracis-33576	53	3	enhances	enhance	VERB
bracis-33576	53	4	clustering	cluster	VERB
bracis-33576	53	5	by	by	ADP
bracis-33576	53	6	utilizing	utilize	VERB
bracis-33576	53	7	llms	llm	NOUN
bracis-33576	53	8	to	to	PART
bracis-33576	53	9	refine	refine	VERB
bracis-33576	53	10	clustering	clustering	ADJ
bracis-33576	53	11	perspectives	perspective	NOUN
bracis-33576	53	12	and	and	CCONJ
bracis-33576	53	13	granularity	granularity	NOUN
bracis-33576	53	14	through	through	ADP
bracis-33576	53	15	two	two	NUM
bracis-33576	53	16	stages	stage	NOUN
bracis-33576	53	17	:	:	PUNCT
bracis-33576	53	18	a	a	DET
bracis-33576	53	19	triplet	triplet	NOUN
bracis-33576	53	20	task	task	NOUN
bracis-33576	53	21	for	for	ADP
bracis-33576	53	22	fine	fine	ADV
bracis-33576	53	23	-	-	PUNCT
bracis-33576	53	24	tuning	tune	VERB
bracis-33576	53	25	embedders	embedder	NOUN
bracis-33576	53	26	based	base	VERB
bracis-33576	53	27	on	on	ADP
bracis-33576	53	28	user	user	NOUN
bracis-33576	53	29	preferences	preference	NOUN
bracis-33576	53	30	and	and	CCONJ
bracis-33576	53	31	a	a	DET
bracis-33576	53	32	pairwise	pairwise	NOUN
bracis-33576	53	33	task	task	NOUN
bracis-33576	53	34	for	for	ADP
bracis-33576	53	35	determining	determine	VERB
bracis-33576	53	36	cluster	cluster	NOUN
bracis-33576	53	37	granularity	granularity	NOUN
bracis-33576	53	38	.	.	PUNCT
bracis-33576	54	1	extensive	extensive	ADJ
bracis-33576	54	2	experiments	experiment	NOUN
bracis-33576	54	3	on	on	ADP
bracis-33576	54	4	fourteen	fourteen	NUM
bracis-33576	54	5	datasets	dataset	NOUN
bracis-33576	54	6	demonstrated	demonstrate	VERB
bracis-33576	54	7	that	that	SCONJ
bracis-33576	54	8	clusterllm	clusterllm	NOUN
bracis-33576	54	9	consistently	consistently	ADV
bracis-33576	54	10	improves	improve	VERB
bracis-33576	54	11	clustering	cluster	VERB
bracis-33576	54	12	quality	quality	NOUN
bracis-33576	54	13	and	and	CCONJ
bracis-33576	54	14	is	be	AUX
bracis-33576	54	15	cost	cost	NOUN
bracis-33576	54	16	-	-	PUNCT
bracis-33576	54	17	effective	effective	ADJ
bracis-33576	54	18	,	,	PUNCT
bracis-33576	54	19	outperforming	outperform	VERB
bracis-33576	54	20	traditional	traditional	ADJ
bracis-33576	54	21	clustering	clustering	ADJ
bracis-33576	54	22	methods	method	NOUN
bracis-33576	54	23	.	.	PUNCT
bracis-33576	55	1	both	both	DET
bracis-33576	55	2	werd	werd	NOUN
bracis-33576	55	3	[	[	X
bracis-33576	55	4	24	24	NUM
bracis-33576	55	5	]	]	PUNCT
bracis-33576	55	6	and	and	CCONJ
bracis-33576	55	7	clusterllm	clusterllm	VERB
bracis-33576	56	1	[	[	X
bracis-33576	56	2	23	23	NUM
bracis-33576	56	3	]	]	PUNCT
bracis-33576	56	4	presented	present	VERB
bracis-33576	56	5	limitations	limitation	NOUN
bracis-33576	56	6	compared	compare	VERB
bracis-33576	56	7	to	to	ADP
bracis-33576	56	8	the	the	DET
bracis-33576	56	9	erasmo	erasmo	ADJ
bracis-33576	56	10	framework	framework	NOUN
bracis-33576	56	11	(	(	PUNCT
bracis-33576	56	12	our	our	PRON
bracis-33576	56	13	proposal	proposal	NOUN
bracis-33576	56	14	)	)	PUNCT
bracis-33576	56	15	.	.	PUNCT
bracis-33576	57	1	werd	werd	NOUN
bracis-33576	57	2	might	might	AUX
bracis-33576	57	3	not	not	PART
bracis-33576	57	4	fully	fully	ADV
bracis-33576	57	5	capture	capture	VERB
bracis-33576	57	6	the	the	DET
bracis-33576	57	7	contextual	contextual	ADJ
bracis-33576	57	8	nuances	nuance	NOUN
bracis-33576	57	9	of	of	ADP
bracis-33576	57	10	each	each	DET
bracis-33576	57	11	dataset	dataset	NOUN
bracis-33576	57	12	due	due	ADP
bracis-33576	57	13	to	to	ADP
bracis-33576	57	14	its	its	PRON
bracis-33576	57	15	focus	focus	NOUN
bracis-33576	57	16	on	on	ADP
bracis-33576	57	17	dimensionality	dimensionality	NOUN
bracis-33576	57	18	reduction	reduction	NOUN
bracis-33576	57	19	techniques	technique	NOUN
bracis-33576	57	20	.	.	PUNCT
bracis-33576	58	1	at	at	ADP
bracis-33576	58	2	the	the	DET
bracis-33576	58	3	same	same	ADJ
bracis-33576	58	4	time	time	NOUN
bracis-33576	58	5	,	,	PUNCT
bracis-33576	58	6	clusterllm	clusterllm	NOUN
bracis-33576	58	7	’s	’s	PART
bracis-33576	58	8	reliance	reliance	NOUN
bracis-33576	58	9	on	on	ADP
bracis-33576	58	10	general	general	ADJ
bracis-33576	58	11	-	-	PUNCT
bracis-33576	58	12	purpose	purpose	NOUN
bracis-33576	58	13	llms	llm	NOUN
bracis-33576	58	14	for	for	ADP
bracis-33576	58	15	guidance	guidance	NOUN
bracis-33576	58	16	may	may	AUX
bracis-33576	58	17	overlook	overlook	VERB
bracis-33576	58	18	specific	specific	ADJ
bracis-33576	58	19	dataset	dataset	NOUN
bracis-33576	58	20	characteristics	characteristic	NOUN
bracis-33576	58	21	.	.	PUNCT
bracis-33576	59	1	a	a	DET
bracis-33576	59	2	method	method	NOUN
bracis-33576	59	3	demonstrating	demonstrating	NOUN
bracis-33576	59	4	that	that	PRON
bracis-33576	59	5	llms	llm	NOUN
bracis-33576	59	6	enables	enable	VERB
bracis-33576	59	7	few	few	ADJ
bracis-33576	59	8	-	-	PUNCT
bracis-33576	59	9	short	short	ADJ
bracis-33576	59	10	learning	learning	NOUN
bracis-33576	59	11	applied	apply	VERB
bracis-33576	59	12	to	to	ADP
bracis-33576	59	13	clustering	clustering	ADJ
bracis-33576	59	14	tasks	task	NOUN
bracis-33576	59	15	was	be	AUX
bracis-33576	59	16	proposed	propose	VERB
bracis-33576	59	17	in	in	ADP
bracis-33576	59	18	[	[	X
bracis-33576	59	19	21	21	NUM
bracis-33576	59	20	]	]	PUNCT
bracis-33576	59	21	.	.	PUNCT
bracis-33576	60	1	their	their	PRON
bracis-33576	60	2	study	study	NOUN
bracis-33576	60	3	showed	show	VERB
bracis-33576	60	4	how	how	SCONJ
bracis-33576	60	5	llms	llm	NOUN
bracis-33576	60	6	can	can	AUX
bracis-33576	60	7	perform	perform	VERB
bracis-33576	60	8	clustering	clustering	ADJ
bracis-33576	60	9	tasks	task	NOUN
bracis-33576	60	10	with	with	ADP
bracis-33576	60	11	minimal	minimal	ADJ
bracis-33576	60	12	labeled	label	VERB
bracis-33576	60	13	data	datum	NOUN
bracis-33576	60	14	by	by	ADP
bracis-33576	60	15	leveraging	leverage	VERB
bracis-33576	60	16	their	their	PRON
bracis-33576	60	17	extensive	extensive	ADJ
bracis-33576	60	18	pretraining	pretraining	NOUN
bracis-33576	60	19	,	,	PUNCT
bracis-33576	60	20	significantly	significantly	ADV
bracis-33576	60	21	reducing	reduce	VERB
bracis-33576	60	22	the	the	DET
bracis-33576	60	23	need	need	NOUN
bracis-33576	60	24	for	for	ADP
bracis-33576	60	25	large	large	ADJ
bracis-33576	60	26	annotated	annotate	VERB
bracis-33576	60	27	datasets	dataset	NOUN
bracis-33576	60	28	and	and	CCONJ
bracis-33576	60	29	achieving	achieve	VERB
bracis-33576	60	30	reasonable	reasonable	ADJ
bracis-33576	60	31	clustering	clustering	ADJ
bracis-33576	60	32	performance	performance	NOUN
bracis-33576	60	33	with	with	ADP
bracis-33576	60	34	few	few	ADJ
bracis-33576	60	35	-	-	PUNCT
bracis-33576	60	36	shot	shot	NOUN
bracis-33576	60	37	learning	learning	NOUN
bracis-33576	60	38	.	.	PUNCT
bracis-33576	61	1	similarly	similarly	ADV
bracis-33576	61	2	,	,	PUNCT
bracis-33576	61	3	tipirneni	tipirneni	PROPN
bracis-33576	61	4	et	et	PROPN
bracis-33576	61	5	al	al	PROPN
bracis-33576	61	6	.	.	PUNCT
bracis-33576	62	1	[	[	X
bracis-33576	62	2	18	18	NUM
bracis-33576	62	3	]	]	PUNCT
bracis-33576	62	4	explored	explore	VERB
bracis-33576	62	5	context	context	NOUN
bracis-33576	62	6	-	-	PUNCT
bracis-33576	62	7	aware	aware	ADJ
bracis-33576	62	8	clustering	clustering	NOUN
bracis-33576	62	9	using	use	VERB
bracis-33576	62	10	llms	llm	NOUN
bracis-33576	62	11	,	,	PUNCT
bracis-33576	62	12	highlighting	highlight	VERB
bracis-33576	62	13	how	how	SCONJ
bracis-33576	62	14	these	these	DET
bracis-33576	62	15	models	model	NOUN
bracis-33576	62	16	can	can	AUX
bracis-33576	62	17	utilize	utilize	VERB
bracis-33576	62	18	contextual	contextual	ADJ
bracis-33576	62	19	information	information	NOUN
bracis-33576	62	20	to	to	PART
bracis-33576	62	21	enhance	enhance	VERB
bracis-33576	62	22	clustering	clustering	ADJ
bracis-33576	62	23	accuracy	accuracy	NOUN
bracis-33576	62	24	.	.	PUNCT
bracis-33576	63	1	both	both	DET
bracis-33576	63	2	methods	method	NOUN
bracis-33576	63	3	,	,	PUNCT
bracis-33576	63	4	however	however	ADV
bracis-33576	63	5	,	,	PUNCT
bracis-33576	63	6	may	may	AUX
bracis-33576	63	7	not	not	PART
bracis-33576	63	8	fully	fully	ADV
bracis-33576	63	9	leverage	leverage	VERB
bracis-33576	63	10	the	the	DET
bracis-33576	63	11	dataset	dataset	NOUN
bracis-33576	63	12	-	-	PUNCT
bracis-33576	63	13	specific	specific	ADJ
bracis-33576	63	14	nuances	nuance	NOUN
bracis-33576	63	15	as	as	ADV
bracis-33576	63	16	effectively	effectively	ADV
bracis-33576	63	17	as	as	ADP
bracis-33576	63	18	erasmo	erasmo	NOUN
bracis-33576	63	19	because	because	SCONJ
bracis-33576	63	20	we	we	PRON
bracis-33576	63	21	employ	employ	VERB
bracis-33576	63	22	a	a	DET
bracis-33576	63	23	fine	fine	ADV
bracis-33576	63	24	-	-	PUNCT
bracis-33576	63	25	tuning	tune	VERB
bracis-33576	63	26	step	step	NOUN
bracis-33576	63	27	,	,	PUNCT
bracis-33576	63	28	allowing	allow	VERB
bracis-33576	63	29	the	the	DET
bracis-33576	63	30	model	model	NOUN
bracis-33576	63	31	to	to	PART
bracis-33576	63	32	capture	capture	VERB
bracis-33576	63	33	better	well	ADV
bracis-33576	63	34	and	and	CCONJ
bracis-33576	63	35	utilize	utilize	VERB
bracis-33576	63	36	dataset	dataset	NOUN
bracis-33576	63	37	-	-	PUNCT
bracis-33576	63	38	specific	specific	ADJ
bracis-33576	63	39	details	detail	NOUN
bracis-33576	63	40	,	,	PUNCT
bracis-33576	63	41	leading	lead	VERB
bracis-33576	63	42	to	to	ADP
bracis-33576	63	43	more	more	ADV
bracis-33576	63	44	accurate	accurate	ADJ
bracis-33576	63	45	and	and	CCONJ
bracis-33576	63	46	reliable	reliable	ADJ
bracis-33576	63	47	clustering	clustering	NOUN
bracis-33576	63	48	results	result	NOUN
bracis-33576	63	49	.	.	PUNCT
bracis-33576	64	1	tissera	tissera	ADV
bracis-33576	64	2	,	,	PUNCT
bracis-33576	64	3	asanka	asanka	NOUN
bracis-33576	64	4	,	,	PUNCT
bracis-33576	64	5	&	&	CCONJ
bracis-33576	64	6	rajapakse	rajapakse	NOUN
bracis-33576	64	7	developed	develop	VERB
bracis-33576	64	8	 	 	SPACE
bracis-33576	65	1	[	[	X
bracis-33576	65	2	19	19	NUM
bracis-33576	65	3	]	]	PUNCT
bracis-33576	65	4	an	an	DET
bracis-33576	65	5	approach	approach	NOUN
bracis-33576	65	6	to	to	ADP
bracis-33576	65	7	enhancing	enhance	VERB
bracis-33576	65	8	customer	customer	NOUN
bracis-33576	65	9	segmentation	segmentation	NOUN
bracis-33576	65	10	using	use	VERB
bracis-33576	65	11	llms	llm	NOUN
bracis-33576	65	12	and	and	CCONJ
bracis-33576	65	13	dice	dice	NOUN
bracis-33576	65	14	.	.	PUNCT
bracis-33576	66	1	their	their	PRON
bracis-33576	66	2	method	method	NOUN
bracis-33576	66	3	combined	combine	VERB
bracis-33576	66	4	llms	llm	NOUN
bracis-33576	66	5	with	with	ADP
bracis-33576	66	6	dice	dice	NOUN
bracis-33576	66	7	to	to	PART
bracis-33576	66	8	generate	generate	VERB
bracis-33576	66	9	consistent	consistent	ADJ
bracis-33576	66	10	and	and	CCONJ
bracis-33576	66	11	deterministic	deterministic	ADJ
bracis-33576	66	12	embeddings	embedding	NOUN
bracis-33576	66	13	across	across	ADP
bracis-33576	66	14	different	different	ADJ
bracis-33576	66	15	datasets	dataset	NOUN
bracis-33576	66	16	,	,	PUNCT
bracis-33576	66	17	improving	improve	VERB
bracis-33576	66	18	segmentation	segmentation	NOUN
bracis-33576	66	19	accuracy	accuracy	NOUN
bracis-33576	66	20	and	and	CCONJ
bracis-33576	66	21	robustness	robustness	NOUN
bracis-33576	66	22	.	.	PUNCT
bracis-33576	67	1	their	their	PRON
bracis-33576	67	2	approach	approach	NOUN
bracis-33576	67	3	may	may	AUX
bracis-33576	67	4	not	not	PART
bracis-33576	67	5	fully	fully	ADV
bracis-33576	67	6	leverage	leverage	VERB
bracis-33576	67	7	the	the	DET
bracis-33576	67	8	context	context	NOUN
bracis-33576	67	9	-	-	PUNCT
bracis-33576	67	10	specific	specific	ADJ
bracis-33576	67	11	nuances	nuance	NOUN
bracis-33576	67	12	of	of	ADP
bracis-33576	67	13	each	each	DET
bracis-33576	67	14	dataset	dataset	VERB
bracis-33576	67	15	as	as	ADV
bracis-33576	67	16	effectively	effectively	ADV
bracis-33576	67	17	as	as	ADP
bracis-33576	67	18	erasmo	erasmo	PROPN
bracis-33576	67	19	.	.	PUNCT
bracis-33576	68	1	our	our	PRON
bracis-33576	68	2	fine	fine	ADV
bracis-33576	68	3	-	-	PUNCT
bracis-33576	68	4	tuning	tune	VERB
bracis-33576	68	5	process	process	NOUN
bracis-33576	68	6	proposal	proposal	NOUN
bracis-33576	68	7	allows	allow	VERB
bracis-33576	68	8	it	it	PRON
bracis-33576	68	9	to	to	PART
bracis-33576	68	10	adapt	adapt	VERB
bracis-33576	68	11	to	to	ADP
bracis-33576	68	12	the	the	DET
bracis-33576	68	13	unique	unique	ADJ
bracis-33576	68	14	characteristics	characteristic	NOUN
bracis-33576	68	15	of	of	ADP
bracis-33576	68	16	the	the	DET
bracis-33576	68	17	input	input	NOUN
bracis-33576	68	18	data	datum	NOUN
bracis-33576	68	19	,	,	PUNCT
bracis-33576	68	20	providing	provide	VERB
bracis-33576	68	21	more	more	ADV
bracis-33576	68	22	contextually	contextually	ADV
bracis-33576	68	23	rich	rich	ADJ
bracis-33576	68	24	and	and	CCONJ
bracis-33576	68	25	detailed	detailed	ADJ
bracis-33576	68	26	embeddings	embedding	NOUN
bracis-33576	68	27	that	that	PRON
bracis-33576	68	28	can	can	AUX
bracis-33576	68	29	result	result	VERB
bracis-33576	68	30	in	in	ADP
bracis-33576	68	31	more	more	ADV
bracis-33576	68	32	precise	precise	ADJ
bracis-33576	68	33	and	and	CCONJ
bracis-33576	68	34	meaningful	meaningful	ADJ
bracis-33576	68	35	clusters	cluster	NOUN
bracis-33576	68	36	.	.	PUNCT
bracis-33576	69	1	a	a	DET
bracis-33576	69	2	comparative	comparative	ADJ
bracis-33576	69	3	analysis	analysis	NOUN
bracis-33576	69	4	of	of	ADP
bracis-33576	69	5	llm	llm	PROPN
bracis-33576	69	6	embeddings	embedding	NOUN
bracis-33576	69	7	for	for	ADP
bracis-33576	69	8	effective	effective	ADJ
bracis-33576	69	9	clustering	clustering	NOUN
bracis-33576	69	10	was	be	AUX
bracis-33576	69	11	explored	explore	VERB
bracis-33576	69	12	in	in	ADP
bracis-33576	69	13	 	 	SPACE
bracis-33576	69	14	[	[	X
bracis-33576	69	15	6	6	NUM
bracis-33576	69	16	]	]	PUNCT
bracis-33576	69	17	.	.	PUNCT
bracis-33576	70	1	as	as	ADP
bracis-33576	70	2	an	an	DET
bracis-33576	70	3	extension	extension	NOUN
bracis-33576	70	4	,	,	PUNCT
bracis-33576	70	5	the	the	DET
bracis-33576	70	6	study	study	NOUN
bracis-33576	70	7	on	on	ADP
bracis-33576	70	8	text	text	NOUN
bracis-33576	70	9	clustering	cluster	VERB
bracis-33576	70	10	with	with	ADP
bracis-33576	70	11	llm	llm	PROPN
bracis-33576	70	12	embeddings	embedding	NOUN
bracis-33576	70	13	 	 	SPACE
bracis-33576	70	14	[	[	X
bracis-33576	70	15	14	14	NUM
bracis-33576	70	16	]	]	PUNCT
bracis-33576	70	17	delves	delve	VERB
bracis-33576	70	18	deeper	deeply	ADV
bracis-33576	70	19	,	,	PUNCT
bracis-33576	70	20	exploring	explore	VERB
bracis-33576	70	21	additional	additional	ADJ
bracis-33576	70	22	models	model	NOUN
bracis-33576	70	23	and	and	CCONJ
bracis-33576	70	24	datasets	dataset	NOUN
bracis-33576	70	25	to	to	PART
bracis-33576	70	26	demonstrate	demonstrate	VERB
bracis-33576	70	27	improvements	improvement	NOUN
bracis-33576	70	28	in	in	ADP
bracis-33576	70	29	text	text	NOUN
bracis-33576	70	30	data	datum	NOUN
bracis-33576	70	31	clustering	cluster	VERB
bracis-33576	70	32	.	.	PUNCT
bracis-33576	71	1	while	while	SCONJ
bracis-33576	71	2	existing	exist	VERB
bracis-33576	71	3	approaches	approach	NOUN
bracis-33576	71	4	effectively	effectively	ADV
bracis-33576	71	5	capture	capture	VERB
bracis-33576	71	6	complex	complex	ADJ
bracis-33576	71	7	semantic	semantic	ADJ
bracis-33576	71	8	relationships	relationship	NOUN
bracis-33576	71	9	and	and	CCONJ
bracis-33576	71	10	handle	handle	VERB
bracis-33576	71	11	categorical	categorical	ADJ
bracis-33576	71	12	,	,	PUNCT
bracis-33576	71	13	numerical	numerical	ADJ
bracis-33576	71	14	,	,	PUNCT
bracis-33576	71	15	and	and	CCONJ
bracis-33576	71	16	textual	textual	ADJ
bracis-33576	71	17	data	datum	NOUN
bracis-33576	71	18	,	,	PUNCT
bracis-33576	71	19	they	they	PRON
bracis-33576	71	20	lack	lack	VERB
bracis-33576	71	21	the	the	DET
bracis-33576	71	22	fine	fine	ADV
bracis-33576	71	23	-	-	PUNCT
bracis-33576	71	24	tuning	tune	VERB
bracis-33576	71	25	specificity	specificity	NOUN
bracis-33576	71	26	of	of	ADP
bracis-33576	71	27	erasmo	erasmo	PROPN
bracis-33576	71	28	,	,	PUNCT
bracis-33576	71	29	our	our	PRON
bracis-33576	71	30	key	key	ADJ
bracis-33576	71	31	originality	originality	NOUN
bracis-33576	71	32	aspect	aspect	NOUN
bracis-33576	71	33	.	.	PUNCT
bracis-33576	72	1	erasmo	erasmo	PROPN
bracis-33576	72	2	’s	’s	PART
bracis-33576	72	3	tailored	tailor	VERB
bracis-33576	72	4	embeddings	embedding	NOUN
bracis-33576	72	5	for	for	ADP
bracis-33576	72	6	tabular	tabular	NOUN
bracis-33576	72	7	datasets	dataset	NOUN
bracis-33576	72	8	and	and	CCONJ
bracis-33576	72	9	integration	integration	NOUN
bracis-33576	72	10	of	of	ADP
bracis-33576	72	11	feature	feature	NOUN
bracis-33576	72	12	order	order	NOUN
bracis-33576	72	13	permutation	permutation	NOUN
bracis-33576	72	14	provide	provide	VERB
bracis-33576	72	15	more	more	ADV
bracis-33576	72	16	precise	precise	ADJ
bracis-33576	72	17	and	and	CCONJ
bracis-33576	72	18	contextually	contextually	ADV
bracis-33576	72	19	relevant	relevant	ADJ
bracis-33576	72	20	clusters	cluster	NOUN
bracis-33576	72	21	,	,	PUNCT
bracis-33576	72	22	offering	offer	VERB
bracis-33576	72	23	superior	superior	ADJ
bracis-33576	72	24	versatility	versatility	NOUN
bracis-33576	72	25	and	and	CCONJ
bracis-33576	72	26	robustness	robustness	NOUN
bracis-33576	72	27	in	in	ADP
bracis-33576	72	28	various	various	ADJ
bracis-33576	72	29	clustering	clustering	ADJ
bracis-33576	72	30	applications	application	NOUN
bracis-33576	72	31	.	.	PUNCT
bracis-33576	73	1	3	3	NUM
bracis-33576	73	2	erasmo	erasmo	PROPN
bracis-33576	73	3	this	this	DET
bracis-33576	73	4	section	section	NOUN
bracis-33576	73	5	introduces	introduce	VERB
bracis-33576	73	6	erasmo	erasmo	PROPN
bracis-33576	73	7	,	,	PUNCT
bracis-33576	73	8	our	our	PRON
bracis-33576	73	9	framework	framework	NOUN
bracis-33576	73	10	that	that	PRON
bracis-33576	73	11	leverages	leverage	VERB
bracis-33576	73	12	transformer	transformer	NOUN
bracis-33576	73	13	-	-	PUNCT
bracis-33576	73	14	based	base	VERB
bracis-33576	73	15	language	language	NOUN
bracis-33576	73	16	models	model	NOUN
bracis-33576	73	17	to	to	PART
bracis-33576	73	18	generate	generate	VERB
bracis-33576	73	19	high	high	ADJ
bracis-33576	73	20	-	-	PUNCT
bracis-33576	73	21	quality	quality	NOUN
bracis-33576	73	22	embeddings	embedding	NOUN
bracis-33576	73	23	from	from	ADP
bracis-33576	73	24	tabular	tabular	PROPN
bracis-33576	73	25	data	datum	NOUN
bracis-33576	73	26	.	.	PUNCT
bracis-33576	74	1	these	these	DET
bracis-33576	74	2	embeddings	embedding	NOUN
bracis-33576	74	3	are	be	AUX
bracis-33576	74	4	particularly	particularly	ADV
bracis-33576	74	5	effective	effective	ADJ
bracis-33576	74	6	for	for	ADP
bracis-33576	74	7	clustering	cluster	VERB
bracis-33576	74	8	analysis	analysis	NOUN
bracis-33576	74	9	,	,	PUNCT
bracis-33576	74	10	allowing	allow	VERB
bracis-33576	74	11	for	for	ADP
bracis-33576	74	12	identifying	identify	VERB
bracis-33576	74	13	patterns	pattern	NOUN
bracis-33576	74	14	and	and	CCONJ
bracis-33576	74	15	groupings	grouping	NOUN
bracis-33576	74	16	within	within	ADP
bracis-33576	74	17	the	the	DET
bracis-33576	74	18	data	datum	NOUN
bracis-33576	74	19	that	that	PRON
bracis-33576	74	20	might	might	AUX
bracis-33576	74	21	not	not	PART
bracis-33576	74	22	be	be	AUX
bracis-33576	74	23	immediately	immediately	ADV
bracis-33576	74	24	apparent	apparent	ADJ
bracis-33576	74	25	.	.	PUNCT
bracis-33576	75	1	the	the	DET
bracis-33576	75	2	process	process	NOUN
bracis-33576	75	3	involves	involve	VERB
bracis-33576	75	4	two	two	NUM
bracis-33576	75	5	main	main	ADJ
bracis-33576	75	6	stages	stage	NOUN
bracis-33576	75	7	:	:	PUNCT
bracis-33576	75	8	(	(	PUNCT
bracis-33576	75	9	1	1	X
bracis-33576	75	10	)	)	PUNCT
bracis-33576	75	11	fine	fine	ADV
bracis-33576	75	12	-	-	PUNCT
bracis-33576	75	13	tuning	tune	VERB
bracis-33576	75	14	a	a	DET
bracis-33576	75	15	pretrained	pretraine	VERB
bracis-33576	75	16	llm	llm	NOUN
bracis-33576	75	17	on	on	ADP
bracis-33576	75	18	a	a	DET
bracis-33576	75	19	textually	textually	ADV
bracis-33576	75	20	encoded	encode	VERB
bracis-33576	75	21	tabular	tabular	NOUN
bracis-33576	75	22	dataset	dataset	VERB
bracis-33576	75	23	;	;	PUNCT
bracis-33576	75	24	and	and	CCONJ
bracis-33576	75	25	(	(	PUNCT
bracis-33576	75	26	2	2	X
bracis-33576	75	27	)	)	PUNCT
bracis-33576	75	28	utilizing	utilize	VERB
bracis-33576	75	29	the	the	DET
bracis-33576	75	30	fine	fine	ADV
bracis-33576	75	31	-	-	PUNCT
bracis-33576	75	32	tuned	tune	VERB
bracis-33576	75	33	model	model	NOUN
bracis-33576	75	34	to	to	PART
bracis-33576	75	35	generate	generate	VERB
bracis-33576	75	36	embeddings	embedding	NOUN
bracis-33576	75	37	,	,	PUNCT
bracis-33576	75	38	which	which	PRON
bracis-33576	75	39	are	be	AUX
bracis-33576	75	40	used	use	VERB
bracis-33576	75	41	by	by	ADP
bracis-33576	75	42	a	a	DET
bracis-33576	75	43	clustering	clustering	ADJ
bracis-33576	75	44	algorithm	algorithm	NOUN
bracis-33576	75	45	.	.	PUNCT
bracis-33576	76	1	these	these	DET
bracis-33576	76	2	designed	design	VERB
bracis-33576	76	3	stages	stage	NOUN
bracis-33576	76	4	were	be	AUX
bracis-33576	76	5	inspired	inspire	VERB
bracis-33576	76	6	by	by	ADP
bracis-33576	76	7	 	 	SPACE
bracis-33576	76	8	[	[	X
bracis-33576	76	9	2	2	NUM
bracis-33576	76	10	]	]	PUNCT
bracis-33576	76	11	.	.	PUNCT
bracis-33576	77	1	subsection	subsection	NOUN
bracis-33576	77	2	 	 	SPACE
bracis-33576	77	3	3.1	3.1	NUM
bracis-33576	77	4	details	detail	NOUN
bracis-33576	77	5	the	the	DET
bracis-33576	77	6	fine	fine	ADV
bracis-33576	77	7	-	-	PUNCT
bracis-33576	77	8	tuning	tuning	NOUN
bracis-33576	77	9	phase	phase	NOUN
bracis-33576	77	10	,	,	PUNCT
bracis-33576	77	11	whereas	whereas	SCONJ
bracis-33576	77	12	subsect	subsect	NOUN
bracis-33576	77	13	.	.	PUNCT
bracis-33576	78	1	3.2	3.2	NUM
bracis-33576	78	2	reports	report	NOUN
bracis-33576	78	3	on	on	ADP
bracis-33576	78	4	the	the	DET
bracis-33576	78	5	embedding	embed	VERB
bracis-33576	78	6	generation	generation	NOUN
bracis-33576	78	7	processes	process	NOUN
bracis-33576	78	8	.	.	PUNCT
bracis-33576	79	1	3.1	3.1	NUM
bracis-33576	79	2	phase	phase	NOUN
bracis-33576	79	3	1	1	NUM
bracis-33576	79	4	:	:	PUNCT
bracis-33576	79	5	fine	fine	ADJ
bracis-33576	79	6	-	-	PUNCT
bracis-33576	79	7	tuning	tune	VERB
bracis-33576	79	8	standard	standard	NOUN
bracis-33576	79	9	pretrained	pretraine	VERB
bracis-33576	79	10	generative	generative	ADJ
bracis-33576	79	11	llms	llm	NOUN
bracis-33576	79	12	expect	expect	VERB
bracis-33576	79	13	sequences	sequence	NOUN
bracis-33576	79	14	of	of	ADP
bracis-33576	79	15	words	word	NOUN
bracis-33576	79	16	as	as	ADP
bracis-33576	79	17	inputs	input	NOUN
bracis-33576	79	18	.	.	PUNCT
bracis-33576	80	1	hence	hence	ADV
bracis-33576	80	2	,	,	PUNCT
bracis-33576	80	3	we	we	PRON
bracis-33576	80	4	convert	convert	VERB
bracis-33576	80	5	each	each	DET
bracis-33576	80	6	row	row	NOUN
bracis-33576	80	7	of	of	ADP
bracis-33576	80	8	our	our	PRON
bracis-33576	80	9	dataset	dataset	NOUN
bracis-33576	80	10	into	into	ADP
bracis-33576	80	11	a	a	DET
bracis-33576	80	12	textual	textual	ADJ
bracis-33576	80	13	representation	representation	NOUN
bracis-33576	80	14	to	to	PART
bracis-33576	80	15	apply	apply	VERB
bracis-33576	80	16	an	an	DET
bracis-33576	80	17	llm	llm	NOUN
bracis-33576	80	18	to	to	ADP
bracis-33576	80	19	tabular	tabular	PROPN
bracis-33576	80	20	data	data	PROPN
bracis-33576	80	21	,	,	PUNCT
bracis-33576	80	22	which	which	PRON
bracis-33576	80	23	can	can	AUX
bracis-33576	80	24	contain	contain	VERB
bracis-33576	80	25	categorical	categorical	ADJ
bracis-33576	80	26	,	,	PUNCT
bracis-33576	80	27	numerical	numerical	ADJ
bracis-33576	80	28	,	,	PUNCT
bracis-33576	80	29	and	and	CCONJ
bracis-33576	80	30	textual	textual	ADJ
bracis-33576	80	31	information	information	NOUN
bracis-33576	80	32	.	.	PUNCT
bracis-33576	81	1	definition	definition	NOUN
bracis-33576	81	2	1	1	NUM
bracis-33576	81	3	(	(	PUNCT
bracis-33576	81	4	textual	textual	ADJ
bracis-33576	81	5	converter	converter	NOUN
bracis-33576	81	6	)	)	PUNCT
bracis-33576	81	7	.	.	PUNCT
bracis-33576	82	1	given	give	VERB
bracis-33576	82	2	a	a	DET
bracis-33576	82	3	tabular	tabular	NOUN
bracis-33576	82	4	dataset	dataset	VERB
bracis-33576	82	5	with	with	ADP
bracis-33576	82	6	m	m	PROPN
bracis-33576	82	7	columns	column	NOUN
bracis-33576	82	8	with	with	ADP
bracis-33576	82	9	feature	feature	NOUN
bracis-33576	82	10	names	name	NOUN
bracis-33576	82	11	\(f_1	\(f_1	PROPN
bracis-33576	82	12	,	,	PUNCT
bracis-33576	82	13	f_2	f_2	PROPN
bracis-33576	82	14	,	,	PUNCT
bracis-33576	82	15	\ldots	\ldots	INTJ
bracis-33576	82	16	,	,	PUNCT
bracis-33576	82	17	f_m\	f_m\	PROPN
bracis-33576	82	18	)	)	PUNCT
bracis-33576	82	19	and	and	CCONJ
bracis-33576	82	20	n	n	NUM
bracis-33576	82	21	rows	row	NOUN
bracis-33576	82	22	of	of	ADP
bracis-33576	82	23	samples	sample	NOUN
bracis-33576	82	24	\(s_1	\(s_1	PROPN
bracis-33576	82	25	,	,	PUNCT
bracis-33576	82	26	\ldots	\ldots	INTJ
bracis-33576	82	27	,	,	PUNCT
bracis-33576	82	28	s_n\	s_n\	ADJ
bracis-33576	82	29	)	)	PUNCT
bracis-33576	82	30	,	,	PUNCT
bracis-33576	82	31	let	let	VERB
bracis-33576	82	32	the	the	DET
bracis-33576	82	33	entry	entry	NOUN
bracis-33576	82	34	\(v_{i	\(v_{i	PROPN
bracis-33576	82	35	,	,	PUNCT
bracis-33576	82	36	j}\	j}\	PROPN
bracis-33576	82	37	)	)	PUNCT
bracis-33576	82	38	,	,	PUNCT
bracis-33576	82	39	\(i	\(i	NOUN
bracis-33576	82	40	\in	\in	PROPN
bracis-33576	82	41	\{1	\{1	PROPN
bracis-33576	82	42	,	,	PUNCT
bracis-33576	82	43	\ldots	\ldots	INTJ
bracis-33576	82	44	,	,	PUNCT
bracis-33576	82	45	n\}\	n\}\	NOUN
bracis-33576	82	46	)	)	PUNCT
bracis-33576	82	47	,	,	PUNCT
bracis-33576	82	48	\(j	\(j	VERB
bracis-33576	82	49	\in	\in	PROPN
bracis-33576	82	50	\{1	\{1	NOUN
bracis-33576	82	51	,	,	PUNCT
bracis-33576	82	52	\ldots	\ldots	INTJ
bracis-33576	82	53	,	,	PUNCT
bracis-33576	82	54	m\}\	m\}\	NOUN
bracis-33576	82	55	)	)	PUNCT
bracis-33576	82	56	represent	represent	VERB
bracis-33576	82	57	the	the	DET
bracis-33576	82	58	value	value	NOUN
bracis-33576	82	59	of	of	ADP
bracis-33576	82	60	the	the	DET
bracis-33576	82	61	j	j	PROPN
bracis-33576	82	62	-	-	PUNCT
bracis-33576	82	63	th	th	VERB
bracis-33576	82	64	feature	feature	NOUN
bracis-33576	82	65	of	of	ADP
bracis-33576	82	66	the	the	DET
bracis-33576	82	67	i	i	PROPN
bracis-33576	82	68	-	-	PUNCT
bracis-33576	82	69	th	th	X
bracis-33576	82	70	data	datum	NOUN
bracis-33576	82	71	point	point	NOUN
bracis-33576	82	72	.	.	PUNCT
bracis-33576	83	1	taking	take	VERB
bracis-33576	83	2	the	the	DET
bracis-33576	83	3	feature	feature	NOUN
bracis-33576	83	4	name	name	NOUN
bracis-33576	83	5	and	and	CCONJ
bracis-33576	83	6	value	value	NOUN
bracis-33576	83	7	into	into	ADP
bracis-33576	83	8	account	account	NOUN
bracis-33576	83	9	,	,	PUNCT
bracis-33576	83	10	each	each	DET
bracis-33576	83	11	sample	sample	NOUN
bracis-33576	83	12	\(s_i\	\(s_i\	NOUN
bracis-33576	83	13	)	)	PUNCT
bracis-33576	83	14	of	of	ADP
bracis-33576	83	15	the	the	DET
bracis-33576	83	16	table	table	NOUN
bracis-33576	83	17	is	be	AUX
bracis-33576	83	18	transformed	transform	VERB
bracis-33576	83	19	into	into	ADP
bracis-33576	83	20	a	a	DET
bracis-33576	83	21	textual	textual	ADJ
bracis-33576	83	22	representation	representation	NOUN
bracis-33576	83	23	\(t_i\	\(t_i\	NOUN
bracis-33576	83	24	)	)	PUNCT
bracis-33576	83	25	using	use	VERB
bracis-33576	83	26	the	the	DET
bracis-33576	83	27	following	follow	VERB
bracis-33576	83	28	subject	subject	ADJ
bracis-33576	83	29	-	-	PUNCT
bracis-33576	83	30	predicate	predicate	NOUN
bracis-33576	83	31	-	-	PUNCT
bracis-33576	83	32	object	object	NOUN
bracis-33576	83	33	transformation	transformation	NOUN
bracis-33576	83	34	:	:	PUNCT
bracis-33576	83	35	$	$	SYM
bracis-33576	83	36	$	$	SYM
bracis-33576	83	37	\begin{aligned	\begin{aligne	VERB
bracis-33576	83	38	}	}	PUNCT
bracis-33576	83	39	t_{i	t_{i	PROPN
bracis-33576	83	40	,	,	PUNCT
bracis-33576	83	41	j	j	NOUN
bracis-33576	83	42	}	}	PUNCT
bracis-33576	83	43	&	&	CCONJ
bracis-33576	83	44	=	=	SYM
bracis-33576	84	1	[	[	X
bracis-33576	84	2	f_j	f_j	X
bracis-33576	84	3	,	,	PUNCT
bracis-33576	84	4	\text	\text	ADP
bracis-33576	84	5	{	{	PUNCT
bracis-33576	84	6	`	`	PUNCT
bracis-33576	84	7	`	`	PUNCT
bracis-33576	84	8	is	be	AUX
bracis-33576	84	9	''	''	PUNCT
bracis-33576	84	10	}	}	PUNCT
bracis-33576	84	11	,	,	PUNCT
bracis-33576	84	12	v_{i	v_{i	PROPN
bracis-33576	84	13	,	,	PUNCT
bracis-33576	84	14	j	j	PROPN
bracis-33576	84	15	}	}	PUNCT
bracis-33576	84	16	,	,	PUNCT
bracis-33576	84	17	\text	\text	ADP
bracis-33576	84	18	{	{	PUNCT
bracis-33576	84	19	`	`	PUNCT
bracis-33576	84	20	`	`	PUNCT
bracis-33576	84	21	,	,	PUNCT
bracis-33576	84	22	''	''	PUNCT
bracis-33576	84	23	}	}	PUNCT
bracis-33576	84	24	]	]	PUNCT
bracis-33576	84	25	,	,	PUNCT
bracis-33576	84	26	\quad	\quad	PROPN
bracis-33576	84	27	\forall	\forall	INTJ
bracis-33576	84	28	i	i	PROPN
bracis-33576	84	29	\in	\in	PROPN
bracis-33576	84	30	\{1	\{1	PROPN
bracis-33576	84	31	,	,	PUNCT
bracis-33576	84	32	\ldots	\ldots	INTJ
bracis-33576	84	33	,	,	PUNCT
bracis-33576	84	34	n\	n\	PROPN
bracis-33576	84	35	}	}	PUNCT
bracis-33576	84	36	,	,	PUNCT
bracis-33576	84	37	j	j	PROPN
bracis-33576	84	38	\in	\in	PROPN
bracis-33576	84	39	\{1	\{1	PROPN
bracis-33576	84	40	,	,	PUNCT
bracis-33576	84	41	\ldots	\ldots	INTJ
bracis-33576	84	42	,	,	PUNCT
bracis-33576	84	43	m\	m\	NOUN
bracis-33576	84	44	}	}	PUNCT
bracis-33576	84	45	\end{aligned}$$	\end{aligned}$$	X
bracis-33576	84	46	(	(	PUNCT
bracis-33576	84	47	1a	1a	NUM
bracis-33576	84	48	)	)	PUNCT
bracis-33576	84	49	$	$	SYM
bracis-33576	84	50	$	$	SYM
bracis-33576	84	51	\begin{aligned	\begin{aligne	VERB
bracis-33576	84	52	}	}	PUNCT
bracis-33576	84	53	t_i	t_i	PUNCT
bracis-33576	84	54	&	&	CCONJ
bracis-33576	84	55	=	=	NOUN
bracis-33576	85	1	[	[	X
bracis-33576	85	2	t_{i,1	t_{i,1	X
bracis-33576	85	3	}	}	PUNCT
bracis-33576	85	4	,	,	PUNCT
bracis-33576	85	5	t_{i,2	t_{i,2	ADV
bracis-33576	85	6	}	}	PUNCT
bracis-33576	85	7	,	,	PUNCT
bracis-33576	85	8	\ldots	\ldots	PROPN
bracis-33576	85	9	,	,	PUNCT
bracis-33576	85	10	t_{i	t_{i	PROPN
bracis-33576	85	11	,	,	PUNCT
bracis-33576	85	12	m	m	NOUN
bracis-33576	85	13	}	}	PUNCT
bracis-33576	85	14	]	]	PUNCT
bracis-33576	85	15	,	,	PUNCT
bracis-33576	85	16	\quad	\quad	PROPN
bracis-33576	85	17	\forall	\forall	INTJ
bracis-33576	85	18	i	i	PROPN
bracis-33576	85	19	\in	\in	PROPN
bracis-33576	85	20	\{1	\{1	PROPN
bracis-33576	85	21	,	,	PUNCT
bracis-33576	85	22	\ldots	\ldots	INTJ
bracis-33576	85	23	,	,	PUNCT
bracis-33576	85	24	n\	n\	NOUN
bracis-33576	85	25	}	}	PUNCT
bracis-33576	85	26	,	,	PUNCT
bracis-33576	85	27	\end{aligned}$$	\end{aligned}$$	X
bracis-33576	85	28	(	(	PUNCT
bracis-33576	85	29	1b	1b	NUM
bracis-33576	85	30	)	)	PUNCT
bracis-33576	85	31	where	where	SCONJ
bracis-33576	85	32	\(t_{i	\(t_{i	NOUN
bracis-33576	85	33	,	,	PUNCT
bracis-33576	85	34	j}\	j}\	PROPN
bracis-33576	85	35	)	)	PUNCT
bracis-33576	85	36	,	,	PUNCT
bracis-33576	85	37	the	the	DET
bracis-33576	85	38	textually	textually	ADV
bracis-33576	85	39	encoded	encode	VERB
bracis-33576	85	40	feature	feature	NOUN
bracis-33576	85	41	,	,	PUNCT
bracis-33576	85	42	is	be	AUX
bracis-33576	85	43	a	a	DET
bracis-33576	85	44	clause	clause	NOUN
bracis-33576	85	45	with	with	ADP
bracis-33576	85	46	information	information	NOUN
bracis-33576	85	47	about	about	ADP
bracis-33576	85	48	a	a	DET
bracis-33576	85	49	single	single	ADJ
bracis-33576	85	50	value	value	NOUN
bracis-33576	85	51	and	and	CCONJ
bracis-33576	85	52	its	its	PRON
bracis-33576	85	53	corresponding	corresponding	ADJ
bracis-33576	85	54	feature	feature	NOUN
bracis-33576	85	55	name	name	NOUN
bracis-33576	85	56	,	,	PUNCT
bracis-33576	85	57	and	and	CCONJ
bracis-33576	85	58	\	\	PROPN
bracis-33576	85	59	(	(	PUNCT
bracis-33576	85	60	[	[	PUNCT
bracis-33576	85	61	\cdot	\cdot	NOUN
bracis-33576	85	62	]	]	SYM
bracis-33576	85	63	\	\	NOUN
bracis-33576	85	64	)	)	PUNCT
bracis-33576	85	65	denotes	denote	VERB
bracis-33576	85	66	the	the	DET
bracis-33576	85	67	concatenation	concatenation	NOUN
bracis-33576	85	68	operator	operator	NOUN
bracis-33576	85	69	.	.	PUNCT
bracis-33576	86	1	by	by	ADP
bracis-33576	86	2	transforming	transform	VERB
bracis-33576	86	3	a	a	DET
bracis-33576	86	4	tabular	tabular	NOUN
bracis-33576	86	5	feature	feature	NOUN
bracis-33576	86	6	vector	vector	NOUN
bracis-33576	86	7	into	into	ADP
bracis-33576	86	8	a	a	DET
bracis-33576	86	9	sequence	sequence	NOUN
bracis-33576	86	10	using	use	VERB
bracis-33576	86	11	the	the	DET
bracis-33576	86	12	textual	textual	ADJ
bracis-33576	86	13	subject	subject	NOUN
bracis-33576	86	14	-	-	PUNCT
bracis-33576	86	15	predicate	predicate	NOUN
bracis-33576	86	16	-	-	PUNCT
bracis-33576	86	17	object	object	NOUN
bracis-33576	86	18	encoding	encoding	NOUN
bracis-33576	86	19	scheme	scheme	NOUN
bracis-33576	86	20	,	,	PUNCT
bracis-33576	86	21	pseudo	pseudo	NOUN
bracis-33576	86	22	-	-	ADJ
bracis-33576	86	23	positional	positional	ADJ
bracis-33576	86	24	information	information	NOUN
bracis-33576	86	25	is	be	AUX
bracis-33576	86	26	artificially	artificially	ADV
bracis-33576	86	27	introduced	introduce	VERB
bracis-33576	86	28	into	into	ADP
bracis-33576	86	29	the	the	DET
bracis-33576	86	30	transformed	transform	VERB
bracis-33576	86	31	tabular	tabular	NOUN
bracis-33576	86	32	data	datum	NOUN
bracis-33576	86	33	sample	sample	NOUN
bracis-33576	86	34	.	.	PUNCT
bracis-33576	87	1	however	however	ADV
bracis-33576	87	2	,	,	PUNCT
bracis-33576	87	3	there	there	PRON
bracis-33576	87	4	is	be	VERB
bracis-33576	87	5	no	no	DET
bracis-33576	87	6	spatial	spatial	ADJ
bracis-33576	87	7	ordering	ordering	NOUN
bracis-33576	87	8	relationship	relationship	NOUN
bracis-33576	87	9	between	between	ADP
bracis-33576	87	10	features	feature	NOUN
bracis-33576	87	11	in	in	ADP
bracis-33576	87	12	tabular	tabular	ADJ
bracis-33576	87	13	datasets	dataset	NOUN
bracis-33576	87	14	.	.	PUNCT
bracis-33576	88	1	we	we	PRON
bracis-33576	88	2	randomly	randomly	VERB
bracis-33576	88	3	permute	permute	VERB
bracis-33576	88	4	the	the	DET
bracis-33576	88	5	encoded	encode	VERB
bracis-33576	88	6	short	short	ADJ
bracis-33576	88	7	sentences	sentence	NOUN
bracis-33576	88	8	\(t_{i	\(t_{i	PROPN
bracis-33576	88	9	,	,	PUNCT
bracis-33576	88	10	j}\	j}\	PROPN
bracis-33576	88	11	)	)	PUNCT
bracis-33576	88	12	of	of	ADP
bracis-33576	88	13	the	the	DET
bracis-33576	88	14	full	full	ADJ
bracis-33576	88	15	textual	textual	ADJ
bracis-33576	88	16	representation	representation	NOUN
bracis-33576	88	17	\(t_i\	\(t_i\	NOUN
bracis-33576	88	18	)	)	PUNCT
bracis-33576	88	19	to	to	PART
bracis-33576	88	20	reconstruct	reconstruct	VERB
bracis-33576	88	21	the	the	DET
bracis-33576	88	22	feature	feature	NOUN
bracis-33576	88	23	order	order	NOUN
bracis-33576	88	24	independence	independence	NOUN
bracis-33576	88	25	.	.	PUNCT
bracis-33576	89	1	definition	definition	NOUN
bracis-33576	89	2	2	2	NUM
bracis-33576	89	3	(	(	PUNCT
bracis-33576	89	4	random	random	ADJ
bracis-33576	89	5	feature	feature	NOUN
bracis-33576	89	6	sequence	sequence	NOUN
bracis-33576	89	7	shuffle	shuffle	NOUN
bracis-33576	89	8	)	)	PUNCT
bracis-33576	89	9	.	.	PUNCT
bracis-33576	90	1	let	let	VERB
bracis-33576	90	2	\(t_i	\(t_i	VERB
bracis-33576	90	3	,	,	PUNCT
bracis-33576	90	4	i	i	PRON
bracis-33576	90	5	\in	\in	PROPN
bracis-33576	90	6	\{1	\{1	PROPN
bracis-33576	90	7	,	,	PUNCT
bracis-33576	90	8	\ldots	\ldots	INTJ
bracis-33576	90	9	,	,	PUNCT
bracis-33576	90	10	n\}\	n\}\	NOUN
bracis-33576	90	11	)	)	PUNCT
bracis-33576	90	12	,	,	PUNCT
bracis-33576	90	13	be	be	AUX
bracis-33576	90	14	a	a	DET
bracis-33576	90	15	textual	textual	ADJ
bracis-33576	90	16	representation	representation	NOUN
bracis-33576	90	17	.	.	PUNCT
bracis-33576	91	1	consider	consider	VERB
bracis-33576	91	2	a	a	DET
bracis-33576	91	3	sequence	sequence	NOUN
bracis-33576	91	4	\(k	\(k	NOUN
bracis-33576	91	5	=	=	SYM
bracis-33576	91	6	(	(	PUNCT
bracis-33576	91	7	k_1	k_1	PROPN
bracis-33576	91	8	,	,	PUNCT
bracis-33576	91	9	\ldots	\ldots	PROPN
bracis-33576	91	10	,	,	PUNCT
bracis-33576	91	11	k_m	k_m	NOUN
bracis-33576	91	12	)	)	PUNCT
bracis-33576	92	1	\	\	NOUN
bracis-33576	92	2	)	)	PUNCT
bracis-33576	92	3	that	that	PRON
bracis-33576	92	4	is	be	AUX
bracis-33576	92	5	a	a	DET
bracis-33576	92	6	permutation	permutation	NOUN
bracis-33576	92	7	of	of	ADP
bracis-33576	92	8	the	the	DET
bracis-33576	92	9	sequence	sequence	NOUN
bracis-33576	92	10	of	of	ADP
bracis-33576	92	11	indices	index	NOUN
bracis-33576	92	12	\	\	PROPN
bracis-33576	92	13	(	(	PUNCT
bracis-33576	92	14	(	(	PUNCT
bracis-33576	92	15	1	1	NUM
bracis-33576	92	16	,	,	PUNCT
bracis-33576	92	17	\ldots	\ldots	INTJ
bracis-33576	92	18	,	,	PUNCT
bracis-33576	92	19	m	m	NOUN
bracis-33576	92	20	)	)	PUNCT
bracis-33576	92	21	\	\	NOUN
bracis-33576	92	22	)	)	PUNCT
bracis-33576	92	23	.	.	PUNCT
bracis-33576	93	1	a	a	DET
bracis-33576	93	2	random	random	ADJ
bracis-33576	93	3	feature	feature	NOUN
bracis-33576	93	4	sequence	sequence	NOUN
bracis-33576	93	5	shuffle	shuffle	NOUN
bracis-33576	93	6	is	be	AUX
bracis-33576	93	7	defined	define	VERB
bracis-33576	93	8	as	as	ADP
bracis-33576	93	9	\(t_i(k	\(t_i(k	X
bracis-33576	93	10	)	)	PUNCT
bracis-33576	93	11	=	=	PUNCT
bracis-33576	94	1	[	[	X
bracis-33576	94	2	t_{i	t_{i	NOUN
bracis-33576	94	3	,	,	PUNCT
bracis-33576	94	4	k_1	k_1	PROPN
bracis-33576	94	5	}	}	PUNCT
bracis-33576	94	6	,	,	PUNCT
bracis-33576	94	7	t_{i	t_{i	PROPN
bracis-33576	94	8	,	,	PUNCT
bracis-33576	94	9	k_2	k_2	PROPN
bracis-33576	94	10	}	}	PUNCT
bracis-33576	94	11	,	,	PUNCT
bracis-33576	94	12	\ldots	\ldots	PROPN
bracis-33576	94	13	,	,	PUNCT
bracis-33576	94	14	t_{i	t_{i	PROPN
bracis-33576	94	15	,	,	PUNCT
bracis-33576	94	16	k_m}]\	k_m}]\	NOUN
bracis-33576	94	17	)	)	PUNCT
bracis-33576	94	18	.	.	PUNCT
bracis-33576	95	1	we	we	PRON
bracis-33576	95	2	fine	fine	ADJ
bracis-33576	95	3	-	-	PUNCT
bracis-33576	95	4	tune	tune	NOUN
bracis-33576	95	5	our	our	PRON
bracis-33576	95	6	generative	generative	ADJ
bracis-33576	95	7	language	language	NOUN
bracis-33576	95	8	model	model	NOUN
bracis-33576	95	9	on	on	ADP
bracis-33576	95	10	samples	sample	NOUN
bracis-33576	95	11	without	without	ADP
bracis-33576	95	12	order	order	NOUN
bracis-33576	95	13	dependencies	dependency	NOUN
bracis-33576	95	14	when	when	SCONJ
bracis-33576	95	15	using	use	VERB
bracis-33576	95	16	shuffled	shuffle	VERB
bracis-33576	95	17	orders	order	NOUN
bracis-33576	95	18	of	of	ADP
bracis-33576	95	19	the	the	DET
bracis-33576	95	20	textually	textually	ADV
bracis-33576	95	21	encoded	encode	VERB
bracis-33576	95	22	features	feature	NOUN
bracis-33576	95	23	.	.	PUNCT
bracis-33576	96	1	moreover	moreover	ADV
bracis-33576	96	2	,	,	PUNCT
bracis-33576	96	3	such	such	ADJ
bracis-33576	96	4	permutations	permutation	NOUN
bracis-33576	96	5	are	be	AUX
bracis-33576	96	6	highly	highly	ADV
bracis-33576	96	7	beneficial	beneficial	ADJ
bracis-33576	96	8	as	as	SCONJ
bracis-33576	96	9	they	they	PRON
bracis-33576	96	10	allow	allow	VERB
bracis-33576	96	11	for	for	ADP
bracis-33576	96	12	arbitrary	arbitrary	ADJ
bracis-33576	96	13	conditioning	conditioning	NOUN
bracis-33576	96	14	in	in	ADP
bracis-33576	96	15	tabular	tabular	PROPN
bracis-33576	96	16	data	data	NOUN
bracis-33576	96	17	generation	generation	NOUN
bracis-33576	96	18	.	.	PUNCT
bracis-33576	97	1	in	in	ADP
bracis-33576	97	2	our	our	PRON
bracis-33576	97	3	experiments	experiment	NOUN
bracis-33576	97	4	,	,	PUNCT
bracis-33576	97	5	we	we	PRON
bracis-33576	97	6	refer	refer	VERB
bracis-33576	97	7	to	to	PART
bracis-33576	97	8	erasmobase	erasmobase	VERB
bracis-33576	97	9	as	as	ADP
bracis-33576	97	10	the	the	DET
bracis-33576	97	11	baseline	baseline	PROPN
bracis-33576	97	12	model	model	NOUN
bracis-33576	97	13	,	,	PUNCT
bracis-33576	97	14	utilizing	utilize	VERB
bracis-33576	97	15	only	only	ADV
bracis-33576	97	16	the	the	DET
bracis-33576	97	17	textual	textual	ADJ
bracis-33576	97	18	converter	converter	NOUN
bracis-33576	97	19	and	and	CCONJ
bracis-33576	97	20	random	random	ADJ
bracis-33576	97	21	feature	feature	NOUN
bracis-33576	97	22	sequence	sequence	NOUN
bracis-33576	97	23	shuffle	shuffle	NOUN
bracis-33576	97	24	.	.	PUNCT
bracis-33576	98	1	in	in	ADP
bracis-33576	98	2	addition	addition	NOUN
bracis-33576	98	3	,	,	PUNCT
bracis-33576	98	4	there	there	PRON
bracis-33576	98	5	is	be	VERB
bracis-33576	98	6	evidence	evidence	NOUN
bracis-33576	98	7	that	that	SCONJ
bracis-33576	98	8	verbalizing	verbalize	VERB
bracis-33576	98	9	numerical	numerical	ADJ
bracis-33576	98	10	tokens	token	NOUN
bracis-33576	98	11	can	can	AUX
bracis-33576	98	12	enhance	enhance	VERB
bracis-33576	98	13	effectiveness	effectiveness	NOUN
bracis-33576	98	14	in	in	ADP
bracis-33576	98	15	specific	specific	ADJ
bracis-33576	98	16	scenarios	scenario	NOUN
bracis-33576	98	17	[	[	X
bracis-33576	98	18	8	8	NUM
bracis-33576	98	19	]	]	PUNCT
bracis-33576	98	20	.	.	PUNCT
bracis-33576	99	1	in	in	ADP
bracis-33576	99	2	this	this	DET
bracis-33576	99	3	sense	sense	NOUN
bracis-33576	99	4	,	,	PUNCT
bracis-33576	99	5	we	we	PRON
bracis-33576	99	6	explore	explore	VERB
bracis-33576	99	7	this	this	DET
bracis-33576	99	8	approach	approach	NOUN
bracis-33576	99	9	,	,	PUNCT
bracis-33576	99	10	naming	name	VERB
bracis-33576	99	11	it	it	PRON
bracis-33576	99	12	erasmonv	erasmonv	VERB
bracis-33576	99	13	,	,	PUNCT
bracis-33576	99	14	as	as	SCONJ
bracis-33576	99	15	follows	follow	VERB
bracis-33576	99	16	.	.	PUNCT
bracis-33576	100	1	definition	definition	NOUN
bracis-33576	100	2	3	3	NUM
bracis-33576	100	3	(	(	PUNCT
bracis-33576	100	4	number	number	NOUN
bracis-33576	100	5	verbalizer	verbalizer	NOUN
bracis-33576	100	6	)	)	PUNCT
bracis-33576	100	7	.	.	PUNCT
bracis-33576	101	1	let	let	VERB
bracis-33576	101	2	\(t_i	\(t_i	VERB
bracis-33576	101	3	,	,	PUNCT
bracis-33576	101	4	i	i	PRON
bracis-33576	101	5	\in	\in	PROPN
bracis-33576	101	6	\{1	\{1	PROPN
bracis-33576	101	7	,	,	PUNCT
bracis-33576	101	8	\ldots	\ldots	INTJ
bracis-33576	101	9	,	,	PUNCT
bracis-33576	101	10	n\}\	n\}\	NOUN
bracis-33576	101	11	)	)	PUNCT
bracis-33576	101	12	,	,	PUNCT
bracis-33576	101	13	be	be	AUX
bracis-33576	101	14	a	a	DET
bracis-33576	101	15	textual	textual	ADJ
bracis-33576	101	16	representation	representation	NOUN
bracis-33576	101	17	,	,	PUNCT
bracis-33576	101	18	and	and	CCONJ
bracis-33576	101	19	\(t_{i	\(t_{i	PROPN
bracis-33576	101	20	,	,	PUNCT
bracis-33576	101	21	j}\	j}\	PROPN
bracis-33576	101	22	)	)	PUNCT
bracis-33576	101	23	be	be	VERB
bracis-33576	101	24	the	the	DET
bracis-33576	101	25	set	set	NOUN
bracis-33576	101	26	of	of	ADP
bracis-33576	101	27	words	word	NOUN
bracis-33576	101	28	of	of	ADP
bracis-33576	101	29	the	the	DET
bracis-33576	101	30	j	j	PROPN
bracis-33576	101	31	-	-	PUNCT
bracis-33576	101	32	th	th	VERB
bracis-33576	101	33	feature	feature	NOUN
bracis-33576	101	34	of	of	ADP
bracis-33576	101	35	\(t_i\	\(t_i\	NOUN
bracis-33576	101	36	)	)	PUNCT
bracis-33576	101	37	.	.	PUNCT
bracis-33576	102	1	a	a	DET
bracis-33576	102	2	number	number	NOUN
bracis-33576	102	3	verbalizer	verbalizer	NOUN
bracis-33576	102	4	is	be	AUX
bracis-33576	102	5	a	a	DET
bracis-33576	102	6	function	function	NOUN
bracis-33576	102	7	v	v	NOUN
bracis-33576	102	8	that	that	PRON
bracis-33576	102	9	receives	receive	VERB
bracis-33576	102	10	as	as	ADP
bracis-33576	102	11	input	input	NOUN
bracis-33576	102	12	a	a	DET
bracis-33576	102	13	word	word	NOUN
bracis-33576	102	14	w	w	ADJ
bracis-33576	102	15	of	of	ADP
bracis-33576	102	16	\(t_{i	\(t_{i	PROPN
bracis-33576	102	17	,	,	PUNCT
bracis-33576	102	18	j}\	j}\	PROPN
bracis-33576	102	19	)	)	PUNCT
bracis-33576	102	20	and	and	CCONJ
bracis-33576	102	21	is	be	AUX
bracis-33576	102	22	defined	define	VERB
bracis-33576	102	23	as	as	ADP
bracis-33576	102	24	:	:	PUNCT
bracis-33576	102	25	$	$	SYM
bracis-33576	102	26	$	$	SYM
bracis-33576	102	27	\begin{aligned	\begin{aligne	VERB
bracis-33576	102	28	}	}	PUNCT
bracis-33576	102	29	v(w	v(w	NOUN
bracis-33576	102	30	)	)	PUNCT
bracis-33576	102	31	&	&	CCONJ
bracis-33576	102	32	=	=	SYM
bracis-33576	102	33	{	{	PUNCT
bracis-33576	102	34	\left\	\left\	PROPN
bracis-33576	102	35	{	{	PUNCT
bracis-33576	102	36	\begin{array}{ll	\begin{array}{ll	PROPN
bracis-33576	102	37	}	}	PUNCT
bracis-33576	102	38	w	w	PROPN
bracis-33576	102	39	,	,	PUNCT
bracis-33576	102	40	\text	\text	INTJ
bracis-33576	102	41	{	{	PUNCT
bracis-33576	102	42	if	if	SCONJ
bracis-33576	102	43	}	}	PUNCT
bracis-33576	102	44	w	w	NOUN
bracis-33576	102	45	is	be	AUX
bracis-33576	102	46	not	not	PART
bracis-33576	102	47	numerical,\\	numerical,\\	PROPN
bracis-33576	102	48	\text	\text	ADV
bracis-33576	102	49	{	{	PUNCT
bracis-33576	102	50	verbalized	verbalize	VERB
bracis-33576	102	51	}	}	PUNCT
bracis-33576	102	52	w	w	NOUN
bracis-33576	102	53	otherwise	otherwise	ADV
bracis-33576	102	54	.	.	PUNCT
bracis-33576	103	1	\end{array}\right	\end{array}\right	ADJ
bracis-33576	103	2	.	.	PUNCT
bracis-33576	103	3	}	}	PUNCT
bracis-33576	103	4	\end{aligned}$$	\end{aligned}$$	X
bracis-33576	103	5	(	(	PUNCT
bracis-33576	103	6	2	2	NUM
bracis-33576	103	7	)	)	PUNCT
bracis-33576	103	8	by	by	ADP
bracis-33576	103	9	applying	apply	VERB
bracis-33576	103	10	this	this	DET
bracis-33576	103	11	transformation	transformation	NOUN
bracis-33576	103	12	on	on	ADP
bracis-33576	103	13	every	every	DET
bracis-33576	103	14	token	token	NOUN
bracis-33576	103	15	of	of	ADP
bracis-33576	103	16	every	every	DET
bracis-33576	103	17	textual	textual	ADJ
bracis-33576	103	18	representation	representation	NOUN
bracis-33576	103	19	,	,	PUNCT
bracis-33576	103	20	we	we	PRON
bracis-33576	103	21	ensure	ensure	VERB
bracis-33576	103	22	that	that	SCONJ
bracis-33576	103	23	any	any	DET
bracis-33576	103	24	numerical	numerical	ADJ
bracis-33576	103	25	information	information	NOUN
bracis-33576	103	26	in	in	ADP
bracis-33576	103	27	the	the	DET
bracis-33576	103	28	text	text	NOUN
bracis-33576	103	29	is	be	AUX
bracis-33576	103	30	verbalized	verbalize	VERB
bracis-33576	103	31	.	.	PUNCT
bracis-33576	104	1	in	in	ADP
bracis-33576	104	2	some	some	DET
bracis-33576	104	3	nlp	nlp	NOUN
bracis-33576	104	4	tasks	task	NOUN
bracis-33576	104	5	,	,	PUNCT
bracis-33576	104	6	such	such	ADJ
bracis-33576	104	7	as	as	ADP
bracis-33576	104	8	clustering	cluster	VERB
bracis-33576	104	9	with	with	ADP
bracis-33576	104	10	embeddings	embedding	NOUN
bracis-33576	104	11	,	,	PUNCT
bracis-33576	104	12	sentiment	sentiment	NOUN
bracis-33576	104	13	analysis	analysis	NOUN
bracis-33576	104	14	,	,	PUNCT
bracis-33576	104	15	and	and	CCONJ
bracis-33576	104	16	text	text	NOUN
bracis-33576	104	17	classification	classification	NOUN
bracis-33576	104	18	,	,	PUNCT
bracis-33576	104	19	verbalizing	verbalize	VERB
bracis-33576	104	20	numbers	number	NOUN
bracis-33576	104	21	can	can	AUX
bracis-33576	104	22	improve	improve	VERB
bracis-33576	104	23	the	the	DET
bracis-33576	104	24	model	model	NOUN
bracis-33576	104	25	’s	’s	PART
bracis-33576	104	26	understanding	understanding	NOUN
bracis-33576	104	27	of	of	ADP
bracis-33576	104	28	the	the	DET
bracis-33576	104	29	context	context	NOUN
bracis-33576	104	30	and	and	CCONJ
bracis-33576	104	31	meaning	meaning	NOUN
bracis-33576	104	32	of	of	ADP
bracis-33576	104	33	numerical	numerical	ADJ
bracis-33576	104	34	values	value	NOUN
bracis-33576	104	35	,	,	PUNCT
bracis-33576	104	36	leading	lead	VERB
bracis-33576	104	37	to	to	ADP
bracis-33576	104	38	more	more	ADV
bracis-33576	104	39	accurate	accurate	ADJ
bracis-33576	104	40	and	and	CCONJ
bracis-33576	104	41	meaningful	meaningful	ADJ
bracis-33576	104	42	results	result	NOUN
bracis-33576	104	43	.	.	PUNCT
bracis-33576	105	1	this	this	DET
bracis-33576	105	2	transformation	transformation	NOUN
bracis-33576	105	3	might	might	AUX
bracis-33576	105	4	not	not	PART
bracis-33576	105	5	be	be	AUX
bracis-33576	105	6	beneficial	beneficial	ADJ
bracis-33576	105	7	in	in	ADP
bracis-33576	105	8	some	some	DET
bracis-33576	105	9	cases	case	NOUN
bracis-33576	105	10	,	,	PUNCT
bracis-33576	105	11	depending	depend	VERB
bracis-33576	105	12	on	on	ADP
bracis-33576	105	13	the	the	DET
bracis-33576	105	14	specific	specific	ADJ
bracis-33576	105	15	nature	nature	NOUN
bracis-33576	105	16	of	of	ADP
bracis-33576	105	17	the	the	DET
bracis-33576	105	18	data	datum	NOUN
bracis-33576	105	19	and	and	CCONJ
bracis-33576	105	20	the	the	DET
bracis-33576	105	21	task	task	NOUN
bracis-33576	105	22	at	at	ADP
bracis-33576	105	23	hand	hand	NOUN
bracis-33576	105	24	 	 	SPACE
bracis-33576	106	1	[	[	X
bracis-33576	106	2	8	8	NUM
bracis-33576	106	3	,	,	PUNCT
bracis-33576	106	4	9	9	NUM
bracis-33576	106	5	]	]	PUNCT
bracis-33576	106	6	.	.	PUNCT
bracis-33576	107	1	fine	fine	ADJ
bracis-33576	107	2	-	-	PUNCT
bracis-33576	107	3	tuning	tune	VERB
bracis-33576	107	4	a	a	DET
bracis-33576	107	5	pretrained	pretraine	VERB
bracis-33576	107	6	auto	auto	NOUN
bracis-33576	107	7	-	-	PUNCT
bracis-33576	107	8	regressive	regressive	ADJ
bracis-33576	107	9	language	language	NOUN
bracis-33576	107	10	model	model	NOUN
bracis-33576	107	11	:	:	PUNCT
bracis-33576	107	12	we	we	PRON
bracis-33576	107	13	describe	describe	VERB
bracis-33576	107	14	the	the	DET
bracis-33576	107	15	fine	fine	ADJ
bracis-33576	107	16	-	-	PUNCT
bracis-33576	107	17	tuning	tune	VERB
bracis-33576	107	18	procedure	procedure	NOUN
bracis-33576	107	19	of	of	ADP
bracis-33576	107	20	a	a	DET
bracis-33576	107	21	pretrained	pretraine	VERB
bracis-33576	107	22	llm	llm	NOUN
bracis-33576	107	23	on	on	ADP
bracis-33576	107	24	the	the	DET
bracis-33576	107	25	encoded	encode	VERB
bracis-33576	107	26	tabular	tabular	NOUN
bracis-33576	107	27	data	datum	NOUN
bracis-33576	107	28	for	for	ADP
bracis-33576	107	29	generation	generation	NOUN
bracis-33576	107	30	tasks	task	NOUN
bracis-33576	107	31	.	.	PUNCT
bracis-33576	108	1	we	we	PRON
bracis-33576	108	2	suppose	suppose	VERB
bracis-33576	108	3	a	a	DET
bracis-33576	108	4	textually	textually	ADV
bracis-33576	108	5	encoded	encode	VERB
bracis-33576	108	6	tabular	tabular	NOUN
bracis-33576	108	7	dataset	dataset	VERB
bracis-33576	108	8	\(t	\(t	NOUN
bracis-33576	108	9	=	=	PUNCT
bracis-33576	108	10	\{t_i(k)\}_{i=1	\{t_i(k)\}_{i=1	PROPN
bracis-33576	108	11	,	,	PUNCT
bracis-33576	108	12	\ldots	\ldots	PROPN
bracis-33576	108	13	,	,	PUNCT
bracis-33576	108	14	n}\	n}\	NUM
bracis-33576	108	15	)	)	PUNCT
bracis-33576	108	16	that	that	PRON
bracis-33576	108	17	was	be	AUX
bracis-33576	108	18	transformed	transform	VERB
bracis-33576	108	19	into	into	ADP
bracis-33576	108	20	text	text	NOUN
bracis-33576	108	21	by	by	ADP
bracis-33576	108	22	the	the	DET
bracis-33576	108	23	proposed	propose	VERB
bracis-33576	108	24	encoding	encoding	NOUN
bracis-33576	108	25	scheme	scheme	NOUN
bracis-33576	108	26	.	.	PUNCT
bracis-33576	109	1	let	let	VERB
bracis-33576	109	2	k	k	PRON
bracis-33576	109	3	be	be	AUX
bracis-33576	109	4	a	a	DET
bracis-33576	109	5	randomly	randomly	ADV
bracis-33576	109	6	drawn	draw	VERB
bracis-33576	109	7	permutation	permutation	NOUN
bracis-33576	109	8	,	,	PUNCT
bracis-33576	109	9	and	and	CCONJ
bracis-33576	109	10	n	n	CCONJ
bracis-33576	109	11	denote	denote	VERB
bracis-33576	109	12	the	the	DET
bracis-33576	109	13	number	number	NOUN
bracis-33576	109	14	of	of	ADP
bracis-33576	109	15	rows	row	NOUN
bracis-33576	109	16	.	.	PUNCT
bracis-33576	110	1	based	base	VERB
bracis-33576	110	2	on	on	ADP
bracis-33576	110	3	user	user	NOUN
bracis-33576	110	4	choice	choice	NOUN
bracis-33576	110	5	,	,	PUNCT
bracis-33576	110	6	the	the	DET
bracis-33576	110	7	pipeline	pipeline	NOUN
bracis-33576	110	8	can	can	AUX
bracis-33576	110	9	proceed	proceed	VERB
bracis-33576	110	10	directly	directly	ADV
bracis-33576	110	11	to	to	ADP
bracis-33576	110	12	fine	fine	ADV
bracis-33576	110	13	-	-	PUNCT
bracis-33576	110	14	tuning	tuning	VERB
bracis-33576	110	15	the	the	DET
bracis-33576	110	16	llm	llm	NOUN
bracis-33576	110	17	to	to	PART
bracis-33576	110	18	generate	generate	VERB
bracis-33576	110	19	erasmobase	erasmobase	NOUN
bracis-33576	110	20	,	,	PUNCT
bracis-33576	110	21	or	or	CCONJ
bracis-33576	110	22	it	it	PRON
bracis-33576	110	23	can	can	AUX
bracis-33576	110	24	first	first	ADV
bracis-33576	110	25	apply	apply	VERB
bracis-33576	110	26	a	a	DET
bracis-33576	110	27	number	number	NOUN
bracis-33576	110	28	verbalizer	verbalizer	NOUN
bracis-33576	110	29	to	to	PART
bracis-33576	110	30	convert	convert	VERB
bracis-33576	110	31	numerical	numerical	ADJ
bracis-33576	110	32	tokens	token	NOUN
bracis-33576	110	33	into	into	ADP
bracis-33576	110	34	their	their	PRON
bracis-33576	110	35	verbal	verbal	ADJ
bracis-33576	110	36	representations	representation	NOUN
bracis-33576	110	37	before	before	ADP
bracis-33576	110	38	fine	fine	ADV
bracis-33576	110	39	-	-	PUNCT
bracis-33576	110	40	tuning	tune	VERB
bracis-33576	110	41	the	the	DET
bracis-33576	110	42	llm	llm	NOUN
bracis-33576	110	43	to	to	PART
bracis-33576	110	44	generate	generate	VERB
bracis-33576	110	45	erasmonv	erasmonv	NOUN
bracis-33576	110	46	.	.	PUNCT
bracis-33576	111	1	to	to	PART
bracis-33576	111	2	be	be	AUX
bracis-33576	111	3	processed	process	VERB
bracis-33576	111	4	with	with	ADP
bracis-33576	111	5	an	an	DET
bracis-33576	111	6	llm	llm	NOUN
bracis-33576	111	7	,	,	PUNCT
bracis-33576	111	8	the	the	DET
bracis-33576	111	9	input	input	NOUN
bracis-33576	111	10	sentences	sentence	VERB
bracis-33576	111	11	\(t	\(t	PROPN
bracis-33576	111	12	\in	\in	PROPN
bracis-33576	111	13	t\	t\	PROPN
bracis-33576	111	14	)	)	PUNCT
bracis-33576	111	15	must	must	AUX
bracis-33576	111	16	be	be	AUX
bracis-33576	111	17	encoded	encode	VERB
bracis-33576	111	18	into	into	ADP
bracis-33576	111	19	a	a	DET
bracis-33576	111	20	sequence	sequence	NOUN
bracis-33576	111	21	of	of	ADP
bracis-33576	111	22	tokens	token	NOUN
bracis-33576	111	23	from	from	ADP
bracis-33576	111	24	a	a	DET
bracis-33576	111	25	discrete	discrete	ADJ
bracis-33576	111	26	and	and	CCONJ
bracis-33576	111	27	finite	finite	ADJ
bracis-33576	111	28	vocabulary	vocabulary	PROPN
bracis-33576	111	29	w.	w.	PROPN
bracis-33576	111	30	these	these	DET
bracis-33576	111	31	tokens	token	NOUN
bracis-33576	111	32	can	can	AUX
bracis-33576	111	33	be	be	AUX
bracis-33576	111	34	character	character	NOUN
bracis-33576	111	35	,	,	PUNCT
bracis-33576	111	36	word	word	NOUN
bracis-33576	111	37	,	,	PUNCT
bracis-33576	111	38	or	or	CCONJ
bracis-33576	111	39	subword	subword	PROPN
bracis-33576	111	40	encodings	encoding	NOUN
bracis-33576	111	41	such	such	ADJ
bracis-33576	111	42	as	as	ADP
bracis-33576	111	43	the	the	DET
bracis-33576	111	44	byte	byte	NOUN
bracis-33576	111	45	-	-	PUNCT
bracis-33576	111	46	pair	pair	NOUN
bracis-33576	111	47	-	-	PUNCT
bracis-33576	111	48	encodings	encoding	NOUN
bracis-33576	111	49	(	(	PUNCT
bracis-33576	111	50	bpe	bpe	PROPN
bracis-33576	111	51	)	)	PUNCT
bracis-33576	111	52	.	.	PUNCT
bracis-33576	112	1	thus	thus	ADV
bracis-33576	112	2	,	,	PUNCT
bracis-33576	112	3	\(t	\(t	PROPN
bracis-33576	112	4	\in	\in	PROPN
bracis-33576	112	5	t\	t\	PROPN
bracis-33576	112	6	)	)	PUNCT
bracis-33576	112	7	is	be	AUX
bracis-33576	112	8	represented	represent	VERB
bracis-33576	112	9	by	by	ADP
bracis-33576	112	10	a	a	DET
bracis-33576	112	11	sequence	sequence	NOUN
bracis-33576	112	12	of	of	ADP
bracis-33576	112	13	tokens	token	NOUN
bracis-33576	112	14	\((w_1	\((w_1	NOUN
bracis-33576	112	15	,	,	PUNCT
bracis-33576	112	16	\ldots	\ldot	NOUN
bracis-33576	112	17	,	,	PUNCT
bracis-33576	112	18	w_j	w_j	NUM
bracis-33576	112	19	)	)	PUNCT
bracis-33576	113	1	=	=	SYM
bracis-33576	113	2	\text	\text	ADJ
bracis-33576	113	3	{	{	PUNCT
bracis-33576	113	4	tokenize}(t)\	tokenize}(t)\	NOUN
bracis-33576	113	5	)	)	PUNCT
bracis-33576	113	6	with	with	ADP
bracis-33576	113	7	tokens	token	NOUN
bracis-33576	113	8	\(w_1	\(w_1	NUM
bracis-33576	113	9	,	,	PUNCT
bracis-33576	113	10	\ldots	\ldots	INTJ
bracis-33576	113	11	,	,	PUNCT
bracis-33576	113	12	w_j	w_j	CCONJ
bracis-33576	113	13	\in	\in	PROPN
bracis-33576	113	14	w\	w\	PROPN
bracis-33576	113	15	)	)	PUNCT
bracis-33576	113	16	,	,	PUNCT
bracis-33576	113	17	where	where	SCONJ
bracis-33576	113	18	j	j	PROPN
bracis-33576	113	19	denotes	denote	VERB
bracis-33576	113	20	the	the	DET
bracis-33576	113	21	number	number	NOUN
bracis-33576	113	22	of	of	ADP
bracis-33576	113	23	tokens	token	NOUN
bracis-33576	113	24	required	require	VERB
bracis-33576	113	25	to	to	PART
bracis-33576	113	26	describe	describe	VERB
bracis-33576	113	27	the	the	DET
bracis-33576	113	28	character	character	NOUN
bracis-33576	113	29	sequence	sequence	NOUN
bracis-33576	113	30	t.	t.	PROPN
bracis-33576	113	31	commonly	commonly	ADV
bracis-33576	113	32	,	,	PUNCT
bracis-33576	113	33	the	the	DET
bracis-33576	113	34	probability	probability	NOUN
bracis-33576	113	35	of	of	ADP
bracis-33576	113	36	natural	natural	ADJ
bracis-33576	113	37	-	-	PUNCT
bracis-33576	113	38	language	language	NOUN
bracis-33576	113	39	sequences	sequence	NOUN
bracis-33576	113	40	is	be	AUX
bracis-33576	113	41	factorized	factorize	VERB
bracis-33576	113	42	in	in	ADP
bracis-33576	113	43	an	an	DET
bracis-33576	113	44	auto	auto	NOUN
bracis-33576	113	45	-	-	PUNCT
bracis-33576	113	46	regressive	regressive	ADJ
bracis-33576	113	47	manner	manner	NOUN
bracis-33576	113	48	in	in	ADP
bracis-33576	113	49	llms	llm	NOUN
bracis-33576	113	50	.	.	PUNCT
bracis-33576	114	1	it	it	PRON
bracis-33576	114	2	is	be	AUX
bracis-33576	114	3	represented	represent	VERB
bracis-33576	114	4	as	as	ADP
bracis-33576	114	5	a	a	DET
bracis-33576	114	6	product	product	NOUN
bracis-33576	114	7	of	of	ADP
bracis-33576	114	8	output	output	NOUN
bracis-33576	114	9	probabilities	probability	NOUN
bracis-33576	114	10	conditioned	condition	VERB
bracis-33576	114	11	on	on	ADP
bracis-33576	114	12	previously	previously	ADV
bracis-33576	114	13	observed	observe	VERB
bracis-33576	114	14	tokens	token	NOUN
bracis-33576	114	15	:	:	PUNCT
bracis-33576	114	16	$	$	SYM
bracis-33576	114	17	$	$	SYM
bracis-33576	114	18	\begin{aligned	\begin{aligne	VERB
bracis-33576	114	19	}	}	PUNCT
bracis-33576	114	20	p(t	p(t	NOUN
bracis-33576	114	21	)	)	PUNCT
bracis-33576	114	22	=	=	SYM
bracis-33576	115	1	p(w_1	p(w_1	PROPN
bracis-33576	115	2	,	,	PUNCT
bracis-33576	115	3	\ldots	\ldots	INTJ
bracis-33576	115	4	,	,	PUNCT
bracis-33576	115	5	w_j	w_j	NUM
bracis-33576	115	6	)	)	PUNCT
bracis-33576	115	7	=	=	PUNCT
bracis-33576	116	1	\prod	\prod	PROPN
bracis-33576	116	2	_	_	PROPN
bracis-33576	116	3	{	{	PUNCT
bracis-33576	116	4	k=1}^j	k=1}^j	PUNCT
bracis-33576	116	5	p(w_k	p(w_k	PROPN
bracis-33576	116	6	\mid	\mid	AUX
bracis-33576	116	7	w_1	w_1	VERB
bracis-33576	116	8	,	,	PUNCT
bracis-33576	116	9	\ldots	\ldots	INTJ
bracis-33576	116	10	,	,	PUNCT
bracis-33576	116	11	w_{k-1	w_{k-1	NUM
bracis-33576	116	12	}	}	PUNCT
bracis-33576	116	13	)	)	PUNCT
bracis-33576	116	14	.	.	PUNCT
bracis-33576	117	1	\end{aligned}$$	\end{aligned}$$	PROPN
bracis-33576	117	2	(	(	PUNCT
bracis-33576	117	3	3	3	NUM
bracis-33576	117	4	)	)	PUNCT
bracis-33576	117	5	as	as	ADP
bracis-33576	117	6	a	a	DET
bracis-33576	117	7	result	result	NOUN
bracis-33576	117	8	,	,	PUNCT
bracis-33576	117	9	an	an	DET
bracis-33576	117	10	end	end	NOUN
bracis-33576	117	11	-	-	PUNCT
bracis-33576	117	12	user	user	NOUN
bracis-33576	117	13	can	can	AUX
bracis-33576	117	14	choose	choose	VERB
bracis-33576	117	15	any	any	DET
bracis-33576	117	16	existing	exist	VERB
bracis-33576	117	17	generative	generative	ADJ
bracis-33576	117	18	language	language	NOUN
bracis-33576	117	19	model	model	NOUN
bracis-33576	117	20	for	for	ADP
bracis-33576	117	21	tabular	tabular	NOUN
bracis-33576	117	22	data	datum	NOUN
bracis-33576	117	23	modeling	modeling	NOUN
bracis-33576	117	24	and	and	CCONJ
bracis-33576	117	25	exploit	exploit	VERB
bracis-33576	117	26	the	the	DET
bracis-33576	117	27	vast	vast	ADJ
bracis-33576	117	28	amount	amount	NOUN
bracis-33576	117	29	of	of	ADP
bracis-33576	117	30	contextual	contextual	ADJ
bracis-33576	117	31	knowledge	knowledge	NOUN
bracis-33576	117	32	presented	present	VERB
bracis-33576	117	33	in	in	ADP
bracis-33576	117	34	these	these	DET
bracis-33576	117	35	models	model	NOUN
bracis-33576	117	36	.	.	PUNCT
bracis-33576	118	1	fine	fine	ADJ
bracis-33576	118	2	-	-	PUNCT
bracis-33576	118	3	tuning	tuning	NOUN
bracis-33576	118	4	enables	enable	VERB
bracis-33576	118	5	the	the	DET
bracis-33576	118	6	model	model	NOUN
bracis-33576	118	7	to	to	PART
bracis-33576	118	8	leverage	leverage	VERB
bracis-33576	118	9	this	this	DET
bracis-33576	118	10	contextual	contextual	ADJ
bracis-33576	118	11	information	information	NOUN
bracis-33576	118	12	with	with	ADP
bracis-33576	118	13	the	the	DET
bracis-33576	118	14	feature	feature	NOUN
bracis-33576	118	15	and	and	CCONJ
bracis-33576	118	16	category	category	NOUN
bracis-33576	118	17	names	name	NOUN
bracis-33576	118	18	to	to	PART
bracis-33576	118	19	enhance	enhance	VERB
bracis-33576	118	20	the	the	DET
bracis-33576	118	21	model	model	NOUN
bracis-33576	118	22	’s	’s	PART
bracis-33576	118	23	capabilities	capability	NOUN
bracis-33576	118	24	.	.	PUNCT
bracis-33576	119	1	figure	figure	NOUN
bracis-33576	119	2	 	 	SPACE
bracis-33576	119	3	1	1	NUM
bracis-33576	119	4	presents	present	VERB
bracis-33576	119	5	the	the	DET
bracis-33576	119	6	pipeline	pipeline	NOUN
bracis-33576	119	7	for	for	ADP
bracis-33576	119	8	erasmo	erasmo	NOUN
bracis-33576	119	9	’s	’s	PART
bracis-33576	119	10	fine	fine	ADV
bracis-33576	119	11	-	-	PUNCT
bracis-33576	119	12	tuning	tune	VERB
bracis-33576	119	13	step	step	NOUN
bracis-33576	119	14	.	.	PUNCT
bracis-33576	120	1	3.2	3.2	NUM
bracis-33576	120	2	phase	phase	NOUN
bracis-33576	120	3	2	2	NUM
bracis-33576	120	4	:	:	PUNCT
bracis-33576	120	5	embedding	embed	VERB
bracis-33576	120	6	generation	generation	NOUN
bracis-33576	120	7	and	and	CCONJ
bracis-33576	120	8	 	 	SPACE
bracis-33576	120	9	clustering	cluster	VERB
bracis-33576	120	10	analysis	analysis	NOUN
bracis-33576	120	11	we	we	PRON
bracis-33576	120	12	generate	generate	VERB
bracis-33576	120	13	embeddings	embedding	NOUN
bracis-33576	120	14	from	from	ADP
bracis-33576	120	15	the	the	DET
bracis-33576	120	16	model	model	NOUN
bracis-33576	120	17	after	after	SCONJ
bracis-33576	120	18	fine	fine	ADV
bracis-33576	120	19	-	-	PUNCT
bracis-33576	120	20	tuning	tune	VERB
bracis-33576	120	21	the	the	DET
bracis-33576	120	22	llm	llm	NOUN
bracis-33576	120	23	on	on	ADP
bracis-33576	120	24	the	the	DET
bracis-33576	120	25	textually	textually	ADV
bracis-33576	120	26	encoded	encode	VERB
bracis-33576	120	27	tabular	tabular	NOUN
bracis-33576	120	28	dataset	dataset	VERB
bracis-33576	120	29	.	.	PUNCT
bracis-33576	121	1	these	these	DET
bracis-33576	121	2	embeddings	embedding	NOUN
bracis-33576	121	3	capture	capture	VERB
bracis-33576	121	4	the	the	DET
bracis-33576	121	5	contextual	contextual	ADJ
bracis-33576	121	6	relationships	relationship	NOUN
bracis-33576	121	7	and	and	CCONJ
bracis-33576	121	8	features	feature	NOUN
bracis-33576	121	9	encoded	encode	VERB
bracis-33576	121	10	during	during	ADP
bracis-33576	121	11	the	the	DET
bracis-33576	121	12	training	training	NOUN
bracis-33576	121	13	phase	phase	NOUN
bracis-33576	121	14	.	.	PUNCT
bracis-33576	122	1	we	we	PRON
bracis-33576	122	2	start	start	VERB
bracis-33576	122	3	by	by	ADP
bracis-33576	122	4	feeding	feed	VERB
bracis-33576	122	5	the	the	DET
bracis-33576	122	6	test	test	NOUN
bracis-33576	122	7	dataset	dataset	NOUN
bracis-33576	122	8	,	,	PUNCT
bracis-33576	122	9	transformed	transform	VERB
bracis-33576	122	10	into	into	ADP
bracis-33576	122	11	its	its	PRON
bracis-33576	122	12	textual	textual	ADJ
bracis-33576	122	13	representation	representation	NOUN
bracis-33576	122	14	,	,	PUNCT
bracis-33576	122	15	into	into	ADP
bracis-33576	122	16	the	the	DET
bracis-33576	122	17	fine	fine	ADV
bracis-33576	122	18	-	-	PUNCT
bracis-33576	122	19	tuned	tune	VERB
bracis-33576	122	20	llm	llm	NOUN
bracis-33576	122	21	.	.	PUNCT
bracis-33576	123	1	the	the	DET
bracis-33576	123	2	model	model	NOUN
bracis-33576	123	3	generates	generate	VERB
bracis-33576	123	4	embeddings	embedding	NOUN
bracis-33576	123	5	for	for	ADP
bracis-33576	123	6	each	each	DET
bracis-33576	123	7	input	input	NOUN
bracis-33576	123	8	sequence	sequence	NOUN
bracis-33576	123	9	,	,	PUNCT
bracis-33576	123	10	providing	provide	VERB
bracis-33576	123	11	a	a	DET
bracis-33576	123	12	high	high	ADJ
bracis-33576	123	13	-	-	PUNCT
bracis-33576	123	14	dimensional	dimensional	ADJ
bracis-33576	123	15	representation	representation	NOUN
bracis-33576	123	16	for	for	ADP
bracis-33576	123	17	each	each	DET
bracis-33576	123	18	sample	sample	NOUN
bracis-33576	123	19	.	.	PUNCT
bracis-33576	124	1	this	this	DET
bracis-33576	124	2	process	process	NOUN
bracis-33576	124	3	ensures	ensure	VERB
bracis-33576	124	4	that	that	SCONJ
bracis-33576	124	5	the	the	DET
bracis-33576	124	6	embeddings	embedding	NOUN
bracis-33576	124	7	preserve	preserve	VERB
bracis-33576	124	8	the	the	DET
bracis-33576	124	9	contextual	contextual	ADJ
bracis-33576	124	10	and	and	CCONJ
bracis-33576	124	11	feature	feature	NOUN
bracis-33576	124	12	relationships	relationship	NOUN
bracis-33576	124	13	learned	learn	VERB
bracis-33576	124	14	during	during	ADP
bracis-33576	124	15	fine	fine	ADV
bracis-33576	124	16	-	-	PUNCT
bracis-33576	124	17	tuning	tuning	NOUN
bracis-33576	124	18	.	.	PUNCT
bracis-33576	125	1	depending	depend	VERB
bracis-33576	125	2	on	on	ADP
bracis-33576	125	3	the	the	DET
bracis-33576	125	4	user	user	NOUN
bracis-33576	125	5	’s	’s	PART
bracis-33576	125	6	choice	choice	NOUN
bracis-33576	125	7	in	in	ADP
bracis-33576	125	8	the	the	DET
bracis-33576	125	9	pipeline	pipeline	NOUN
bracis-33576	125	10	,	,	PUNCT
bracis-33576	125	11	the	the	DET
bracis-33576	125	12	embeddings	embedding	NOUN
bracis-33576	125	13	are	be	AUX
bracis-33576	125	14	generated	generate	VERB
bracis-33576	125	15	from	from	ADP
bracis-33576	125	16	either	either	CCONJ
bracis-33576	125	17	erasmobase	erasmobase	NOUN
bracis-33576	125	18	or	or	CCONJ
bracis-33576	125	19	erasmonv	erasmonv	ADJ
bracis-33576	125	20	models	model	NOUN
bracis-33576	125	21	,	,	PUNCT
bracis-33576	125	22	reflecting	reflect	VERB
bracis-33576	125	23	whether	whether	SCONJ
bracis-33576	125	24	the	the	DET
bracis-33576	125	25	number	number	NOUN
bracis-33576	125	26	verbalizer	verbalizer	NOUN
bracis-33576	125	27	step	step	NOUN
bracis-33576	125	28	was	be	AUX
bracis-33576	125	29	applied	apply	VERB
bracis-33576	125	30	.	.	PUNCT
bracis-33576	126	1	to	to	PART
bracis-33576	126	2	generate	generate	VERB
bracis-33576	126	3	these	these	DET
bracis-33576	126	4	embeddings	embedding	NOUN
bracis-33576	126	5	,	,	PUNCT
bracis-33576	126	6	the	the	DET
bracis-33576	126	7	input	input	NOUN
bracis-33576	126	8	sentences	sentence	VERB
bracis-33576	126	9	\(t	\(t	PROPN
bracis-33576	126	10	\in	\in	PROPN
bracis-33576	126	11	t_{test}\	t_{test}\	ADV
bracis-33576	126	12	)	)	PUNCT
bracis-33576	126	13	are	be	AUX
bracis-33576	126	14	encoded	encode	VERB
bracis-33576	126	15	into	into	ADP
bracis-33576	126	16	sequences	sequence	NOUN
bracis-33576	126	17	of	of	ADP
bracis-33576	126	18	tokens	token	NOUN
bracis-33576	126	19	and	and	CCONJ
bracis-33576	126	20	processed	process	VERB
bracis-33576	126	21	by	by	ADP
bracis-33576	126	22	the	the	DET
bracis-33576	126	23	fine	fine	ADV
bracis-33576	126	24	-	-	PUNCT
bracis-33576	126	25	tuned	tune	VERB
bracis-33576	126	26	llm	llm	NOUN
bracis-33576	126	27	.	.	PUNCT
bracis-33576	127	1	the	the	DET
bracis-33576	127	2	embeddings	embedding	NOUN
bracis-33576	127	3	are	be	AUX
bracis-33576	127	4	obtained	obtain	VERB
bracis-33576	127	5	from	from	ADP
bracis-33576	127	6	the	the	DET
bracis-33576	127	7	final	final	ADJ
bracis-33576	127	8	hidden	hide	VERB
bracis-33576	127	9	states	state	NOUN
bracis-33576	127	10	of	of	ADP
bracis-33576	127	11	the	the	DET
bracis-33576	127	12	model	model	NOUN
bracis-33576	127	13	,	,	PUNCT
bracis-33576	127	14	resulting	result	VERB
bracis-33576	127	15	in	in	ADP
bracis-33576	127	16	rich	rich	ADJ
bracis-33576	127	17	and	and	CCONJ
bracis-33576	127	18	informative	informative	ADJ
bracis-33576	127	19	representations	representation	NOUN
bracis-33576	127	20	of	of	ADP
bracis-33576	127	21	the	the	DET
bracis-33576	127	22	data	datum	NOUN
bracis-33576	127	23	.	.	PUNCT
bracis-33576	128	1	these	these	DET
bracis-33576	128	2	embeddings	embedding	NOUN
bracis-33576	128	3	can	can	AUX
bracis-33576	128	4	then	then	ADV
bracis-33576	128	5	be	be	AUX
bracis-33576	128	6	utilized	utilize	VERB
bracis-33576	128	7	for	for	ADP
bracis-33576	128	8	various	various	ADJ
bracis-33576	128	9	downstream	downstream	ADJ
bracis-33576	128	10	tasks	task	NOUN
bracis-33576	128	11	,	,	PUNCT
bracis-33576	128	12	including	include	VERB
bracis-33576	128	13	clustering	cluster	VERB
bracis-33576	128	14	analysis	analysis	NOUN
bracis-33576	128	15	,	,	PUNCT
bracis-33576	128	16	to	to	PART
bracis-33576	128	17	gain	gain	VERB
bracis-33576	128	18	deeper	deep	ADJ
bracis-33576	128	19	insights	insight	NOUN
bracis-33576	128	20	into	into	ADP
bracis-33576	128	21	the	the	DET
bracis-33576	128	22	data	data	NOUN
bracis-33576	128	23	structure	structure	NOUN
bracis-33576	128	24	(	(	PUNCT
bracis-33576	128	25	cf	cf	NOUN
bracis-33576	128	26	.	.	PUNCT
bracis-33576	128	27	figure	figure	NOUN
bracis-33576	128	28	 	 	SPACE
bracis-33576	128	29	2	2	NUM
bracis-33576	128	30	)	)	PUNCT
bracis-33576	128	31	.	.	PUNCT
bracis-33576	129	1	4	4	NUM
bracis-33576	129	2	experimental	experimental	ADJ
bracis-33576	129	3	methodology	methodology	NOUN
bracis-33576	129	4	our	our	PRON
bracis-33576	129	5	experiments	experiment	NOUN
bracis-33576	129	6	evaluated	evaluate	VERB
bracis-33576	129	7	the	the	DET
bracis-33576	129	8	quality	quality	NOUN
bracis-33576	129	9	assessment	assessment	NOUN
bracis-33576	129	10	for	for	ADP
bracis-33576	129	11	the	the	DET
bracis-33576	129	12	best	well	ADV
bracis-33576	129	13	-	-	PUNCT
bracis-33576	129	14	performing	perform	VERB
bracis-33576	129	15	clustering	clustering	ADJ
bracis-33576	129	16	algorithms	algorithm	NOUN
bracis-33576	129	17	for	for	ADP
bracis-33576	129	18	each	each	DET
bracis-33576	129	19	dataset	dataset	NOUN
bracis-33576	129	20	and	and	CCONJ
bracis-33576	129	21	approach	approach	NOUN
bracis-33576	129	22	(	(	PUNCT
bracis-33576	129	23	model	model	NOUN
bracis-33576	129	24	)	)	PUNCT
bracis-33576	129	25	combination	combination	NOUN
bracis-33576	129	26	(	(	PUNCT
bracis-33576	129	27	cf	cf	NOUN
bracis-33576	129	28	.	.	PUNCT
bracis-33576	129	29	table	table	NOUN
bracis-33576	129	30	 	 	SPACE
bracis-33576	129	31	1	1	NUM
bracis-33576	129	32	for	for	ADP
bracis-33576	129	33	the	the	DET
bracis-33576	129	34	obtained	obtain	VERB
bracis-33576	129	35	results	result	NOUN
bracis-33576	129	36	)	)	PUNCT
bracis-33576	129	37	.	.	PUNCT
bracis-33576	130	1	subsection	subsection	NOUN
bracis-33576	130	2	 	 	SPACE
bracis-33576	130	3	4.1	4.1	NUM
bracis-33576	130	4	describes	describe	VERB
bracis-33576	130	5	the	the	DET
bracis-33576	130	6	datasets	dataset	NOUN
bracis-33576	130	7	used	use	VERB
bracis-33576	130	8	for	for	ADP
bracis-33576	130	9	training	training	NOUN
bracis-33576	130	10	and	and	CCONJ
bracis-33576	130	11	testing	testing	NOUN
bracis-33576	130	12	.	.	PUNCT
bracis-33576	131	1	subsection	subsection	NOUN
bracis-33576	131	2	 	 	SPACE
bracis-33576	131	3	4.2	4.2	NUM
bracis-33576	131	4	presents	present	VERB
bracis-33576	131	5	an	an	DET
bracis-33576	131	6	overview	overview	NOUN
bracis-33576	131	7	of	of	ADP
bracis-33576	131	8	the	the	DET
bracis-33576	131	9	clustering	clustering	ADJ
bracis-33576	131	10	algorithms	algorithm	NOUN
bracis-33576	131	11	.	.	PUNCT
bracis-33576	132	1	subsection	subsection	NOUN
bracis-33576	132	2	 	 	SPACE
bracis-33576	132	3	4.3	4.3	NUM
bracis-33576	132	4	describes	describe	VERB
bracis-33576	132	5	the	the	DET
bracis-33576	132	6	approaches	approach	NOUN
bracis-33576	132	7	used	use	VERB
bracis-33576	132	8	for	for	ADP
bracis-33576	132	9	comparison	comparison	NOUN
bracis-33576	132	10	as	as	ADP
bracis-33576	132	11	baselines	baseline	NOUN
bracis-33576	132	12	in	in	ADP
bracis-33576	132	13	our	our	PRON
bracis-33576	132	14	experiments	experiment	NOUN
bracis-33576	132	15	.	.	PUNCT
bracis-33576	133	1	subsection	subsection	NOUN
bracis-33576	133	2	 	 	SPACE
bracis-33576	133	3	4.4	4.4	NUM
bracis-33576	133	4	reports	report	NOUN
bracis-33576	133	5	on	on	ADP
bracis-33576	133	6	the	the	DET
bracis-33576	133	7	evaluation	evaluation	NOUN
bracis-33576	133	8	metrics	metric	NOUN
bracis-33576	133	9	.	.	PUNCT
bracis-33576	134	1	subsection	subsection	NOUN
bracis-33576	134	2	 	 	SPACE
bracis-33576	134	3	4.5	4.5	PROPN
bracis-33576	134	4	presents	present	VERB
bracis-33576	134	5	the	the	DET
bracis-33576	134	6	implementation	implementation	NOUN
bracis-33576	134	7	details	detail	NOUN
bracis-33576	134	8	.	.	PUNCT
bracis-33576	135	1	each	each	DET
bracis-33576	135	2	experimental	experimental	ADJ
bracis-33576	135	3	setup	setup	NOUN
bracis-33576	135	4	for	for	ADP
bracis-33576	135	5	a	a	DET
bracis-33576	135	6	given	give	VERB
bracis-33576	135	7	dataset	dataset	NOUN
bracis-33576	135	8	assessed	assess	VERB
bracis-33576	135	9	the	the	DET
bracis-33576	135	10	different	different	ADJ
bracis-33576	135	11	language	language	NOUN
bracis-33576	135	12	models	model	NOUN
bracis-33576	135	13	fine	fine	ADV
bracis-33576	135	14	-	-	PUNCT
bracis-33576	135	15	tuned	tune	VERB
bracis-33576	135	16	and	and	CCONJ
bracis-33576	135	17	pretrained	pretraine	VERB
bracis-33576	135	18	.	.	PUNCT
bracis-33576	136	1	each	each	DET
bracis-33576	136	2	model	model	NOUN
bracis-33576	136	3	considers	consider	VERB
bracis-33576	136	4	the	the	DET
bracis-33576	136	5	several	several	ADJ
bracis-33576	136	6	clustering	clustering	ADJ
bracis-33576	136	7	algorithms	algorithm	NOUN
bracis-33576	136	8	and	and	CCONJ
bracis-33576	136	9	their	their	PRON
bracis-33576	136	10	configuration	configuration	NOUN
bracis-33576	136	11	.	.	PUNCT
bracis-33576	137	1	fig	fig	NOUN
bracis-33576	137	2	.	.	PUNCT
bracis-33576	138	1	1	1	X
bracis-33576	138	2	.	.	PUNCT
bracis-33576	138	3	the	the	DET
bracis-33576	138	4	erasmo	erasmo	PROPN
bracis-33576	138	5	data	datum	NOUN
bracis-33576	138	6	pipeline	pipeline	NOUN
bracis-33576	138	7	for	for	ADP
bracis-33576	138	8	the	the	DET
bracis-33576	138	9	fine	fine	ADV
bracis-33576	138	10	-	-	PUNCT
bracis-33576	138	11	tuning	tuning	NOUN
bracis-33576	138	12	phase	phase	NOUN
bracis-33576	138	13	.	.	PUNCT
bracis-33576	139	1	first	first	ADV
bracis-33576	139	2	,	,	PUNCT
bracis-33576	139	3	a	a	DET
bracis-33576	139	4	textual	textual	ADJ
bracis-33576	139	5	converter	converter	NOUN
bracis-33576	139	6	step	step	NOUN
bracis-33576	139	7	transforms	transform	VERB
bracis-33576	139	8	tabular	tabular	PROPN
bracis-33576	139	9	data	datum	NOUN
bracis-33576	139	10	into	into	ADP
bracis-33576	139	11	meaningful	meaningful	ADJ
bracis-33576	139	12	text	text	NOUN
bracis-33576	139	13	(	(	PUNCT
bracis-33576	139	14	1	1	NUM
bracis-33576	139	15	)	)	PUNCT
bracis-33576	139	16	.	.	PUNCT
bracis-33576	140	1	next	next	ADV
bracis-33576	140	2	,	,	PUNCT
bracis-33576	140	3	a	a	DET
bracis-33576	140	4	random	random	ADJ
bracis-33576	140	5	feature	feature	NOUN
bracis-33576	140	6	order	order	NOUN
bracis-33576	140	7	permutation	permutation	NOUN
bracis-33576	140	8	step	step	NOUN
bracis-33576	140	9	is	be	AUX
bracis-33576	140	10	applied	apply	VERB
bracis-33576	140	11	(	(	PUNCT
bracis-33576	140	12	2	2	NUM
bracis-33576	140	13	)	)	PUNCT
bracis-33576	140	14	.	.	PUNCT
bracis-33576	141	1	then	then	ADV
bracis-33576	141	2	,	,	PUNCT
bracis-33576	141	3	based	base	VERB
bracis-33576	141	4	on	on	ADP
bracis-33576	141	5	user	user	NOUN
bracis-33576	141	6	choice	choice	NOUN
bracis-33576	141	7	,	,	PUNCT
bracis-33576	141	8	the	the	DET
bracis-33576	141	9	pipeline	pipeline	NOUN
bracis-33576	141	10	diverges	diverge	VERB
bracis-33576	141	11	:	:	PUNCT
bracis-33576	141	12	it	it	PRON
bracis-33576	141	13	can	can	AUX
bracis-33576	141	14	proceed	proceed	VERB
bracis-33576	141	15	directly	directly	ADV
bracis-33576	141	16	to	to	ADP
bracis-33576	141	17	fine	fine	ADV
bracis-33576	141	18	-	-	PUNCT
bracis-33576	141	19	tuning	tuning	VERB
bracis-33576	141	20	a	a	DET
bracis-33576	141	21	llm	llm	NOUN
bracis-33576	141	22	(	(	PUNCT
bracis-33576	141	23	3a	3a	NUM
bracis-33576	141	24	)	)	PUNCT
bracis-33576	141	25	to	to	PART
bracis-33576	141	26	generate	generate	VERB
bracis-33576	141	27	erasmobase	erasmobase	NOUN
bracis-33576	141	28	,	,	PUNCT
bracis-33576	141	29	or	or	CCONJ
bracis-33576	141	30	apply	apply	VERB
bracis-33576	141	31	a	a	DET
bracis-33576	141	32	number	number	NOUN
bracis-33576	141	33	verbalizer	verbalizer	NOUN
bracis-33576	141	34	(	(	PUNCT
bracis-33576	141	35	3b	3b	NUM
bracis-33576	141	36	)	)	PUNCT
bracis-33576	141	37	before	before	ADP
bracis-33576	141	38	fine	fine	ADV
bracis-33576	141	39	-	-	PUNCT
bracis-33576	141	40	tuning	tune	VERB
bracis-33576	141	41	the	the	DET
bracis-33576	141	42	llm	llm	NOUN
bracis-33576	141	43	(	(	PUNCT
bracis-33576	141	44	4b	4b	PROPN
bracis-33576	141	45	)	)	PUNCT
bracis-33576	141	46	to	to	PART
bracis-33576	141	47	generate	generate	VERB
bracis-33576	141	48	erasmonv	erasmonv	NOUN
bracis-33576	141	49	.	.	PUNCT
bracis-33576	142	1	full	full	ADJ
bracis-33576	142	2	size	size	NOUN
bracis-33576	142	3	image	image	NOUN
bracis-33576	142	4	fig	fig	NOUN
bracis-33576	142	5	.	.	PUNCT
bracis-33576	143	1	2	2	X
bracis-33576	143	2	.	.	X
bracis-33576	143	3	the	the	DET
bracis-33576	143	4	erasmo	erasmo	PROPN
bracis-33576	143	5	pipeline	pipeline	NOUN
bracis-33576	143	6	for	for	ADP
bracis-33576	143	7	generating	generate	VERB
bracis-33576	143	8	embeddings	embedding	NOUN
bracis-33576	143	9	and	and	CCONJ
bracis-33576	143	10	cluster	cluster	NOUN
bracis-33576	143	11	analysis	analysis	NOUN
bracis-33576	143	12	.	.	PUNCT
bracis-33576	144	1	the	the	DET
bracis-33576	144	2	input	input	NOUN
bracis-33576	144	3	test	test	NOUN
bracis-33576	144	4	tabular	tabular	PROPN
bracis-33576	144	5	data	datum	NOUN
bracis-33576	144	6	is	be	AUX
bracis-33576	144	7	first	first	ADV
bracis-33576	144	8	transformed	transform	VERB
bracis-33576	144	9	into	into	ADP
bracis-33576	144	10	text	text	NOUN
bracis-33576	144	11	sequences	sequence	NOUN
bracis-33576	144	12	(	(	PUNCT
bracis-33576	144	13	1	1	NUM
bracis-33576	144	14	)	)	PUNCT
bracis-33576	144	15	.	.	PUNCT
bracis-33576	145	1	next	next	ADV
bracis-33576	145	2	,	,	PUNCT
bracis-33576	145	3	a	a	DET
bracis-33576	145	4	random	random	ADJ
bracis-33576	145	5	feature	feature	NOUN
bracis-33576	145	6	order	order	NOUN
bracis-33576	145	7	permutation	permutation	NOUN
bracis-33576	145	8	step	step	NOUN
bracis-33576	145	9	is	be	AUX
bracis-33576	145	10	applied	apply	VERB
bracis-33576	145	11	(	(	PUNCT
bracis-33576	145	12	2	2	NUM
bracis-33576	145	13	)	)	PUNCT
bracis-33576	145	14	.	.	PUNCT
bracis-33576	146	1	for	for	ADP
bracis-33576	146	2	erasmonv	erasmonv	NOUN
bracis-33576	146	3	,	,	PUNCT
bracis-33576	146	4	a	a	DET
bracis-33576	146	5	number	number	NOUN
bracis-33576	146	6	verbalizer	verbalizer	NOUN
bracis-33576	146	7	step	step	NOUN
bracis-33576	146	8	follows	follow	VERB
bracis-33576	146	9	(	(	PUNCT
bracis-33576	146	10	3	3	NUM
bracis-33576	146	11	)	)	PUNCT
bracis-33576	146	12	before	before	ADP
bracis-33576	146	13	processing	process	VERB
bracis-33576	146	14	by	by	ADP
bracis-33576	146	15	the	the	DET
bracis-33576	146	16	fine	fine	ADV
bracis-33576	146	17	-	-	PUNCT
bracis-33576	146	18	tuned	tune	VERB
bracis-33576	146	19	llm	llm	NOUN
bracis-33576	146	20	to	to	PART
bracis-33576	146	21	generate	generate	VERB
bracis-33576	146	22	embeddings	embedding	NOUN
bracis-33576	146	23	(	(	PUNCT
bracis-33576	146	24	4	4	NUM
bracis-33576	146	25	)	)	PUNCT
bracis-33576	146	26	.	.	PUNCT
bracis-33576	147	1	for	for	ADP
bracis-33576	147	2	erasmobase	erasmobase	PROPN
bracis-33576	147	3	,	,	PUNCT
bracis-33576	147	4	the	the	DET
bracis-33576	147	5	pipeline	pipeline	NOUN
bracis-33576	147	6	goes	go	VERB
bracis-33576	147	7	directly	directly	ADV
bracis-33576	147	8	from	from	ADP
bracis-33576	147	9	step	step	NOUN
bracis-33576	147	10	(	(	PUNCT
bracis-33576	147	11	2	2	NUM
bracis-33576	147	12	)	)	PUNCT
bracis-33576	147	13	to	to	PART
bracis-33576	147	14	step	step	VERB
bracis-33576	147	15	(	(	PUNCT
bracis-33576	147	16	4	4	NUM
bracis-33576	147	17	)	)	PUNCT
bracis-33576	147	18	.	.	PUNCT
bracis-33576	148	1	the	the	DET
bracis-33576	148	2	embeddings	embedding	NOUN
bracis-33576	148	3	are	be	AUX
bracis-33576	148	4	subsequently	subsequently	ADV
bracis-33576	148	5	used	use	VERB
bracis-33576	148	6	for	for	ADP
bracis-33576	148	7	clustering	cluster	VERB
bracis-33576	148	8	analysis	analysis	NOUN
bracis-33576	148	9	.	.	PUNCT
bracis-33576	149	1	full	full	ADJ
bracis-33576	149	2	size	size	NOUN
bracis-33576	149	3	image	image	NOUN
bracis-33576	149	4	4.1	4.1	NUM
bracis-33576	149	5	datasets	dataset	NOUN
bracis-33576	149	6	we	we	PRON
bracis-33576	149	7	selected	select	VERB
bracis-33576	149	8	a	a	DET
bracis-33576	149	9	diversified	diversified	ADJ
bracis-33576	149	10	set	set	NOUN
bracis-33576	149	11	of	of	ADP
bracis-33576	149	12	datasets	dataset	NOUN
bracis-33576	149	13	to	to	PART
bracis-33576	149	14	encompass	encompass	VERB
bracis-33576	149	15	a	a	DET
bracis-33576	149	16	variety	variety	NOUN
bracis-33576	149	17	of	of	ADP
bracis-33576	149	18	challenges	challenge	NOUN
bracis-33576	149	19	related	relate	VERB
bracis-33576	149	20	to	to	ADP
bracis-33576	149	21	text	text	NOUN
bracis-33576	149	22	categorization	categorization	NOUN
bracis-33576	149	23	and	and	CCONJ
bracis-33576	149	24	clustering	clustering	NOUN
bracis-33576	149	25	,	,	PUNCT
bracis-33576	149	26	and	and	CCONJ
bracis-33576	149	27	we	we	PRON
bracis-33576	149	28	used	use	VERB
bracis-33576	149	29	them	they	PRON
bracis-33576	149	30	to	to	PART
bracis-33576	149	31	evaluate	evaluate	VERB
bracis-33576	149	32	text	text	NOUN
bracis-33576	149	33	clustering	clustering	ADJ
bracis-33576	149	34	algorithms	algorithm	NOUN
bracis-33576	149	35	.	.	PUNCT
bracis-33576	150	1	banking	banking	NOUN
bracis-33576	150	2	marketing	marketing	NOUN
bracis-33576	150	3	targets	target	NOUN
bracis-33576	150	4	:	:	PUNCT
bracis-33576	150	5	composed	compose	VERB
bracis-33576	150	6	of	of	ADP
bracis-33576	150	7	data	datum	NOUN
bracis-33576	150	8	from	from	ADP
bracis-33576	150	9	direct	direct	ADJ
bracis-33576	150	10	marketing	marketing	NOUN
bracis-33576	150	11	campaigns	campaign	NOUN
bracis-33576	150	12	of	of	ADP
bracis-33576	150	13	a	a	DET
bracis-33576	150	14	banking	banking	NOUN
bracis-33576	150	15	institution	institution	NOUN
bracis-33576	150	16	,	,	PUNCT
bracis-33576	150	17	which	which	PRON
bracis-33576	150	18	includes	include	VERB
bracis-33576	150	19	client	client	NOUN
bracis-33576	150	20	attributes	attribute	NOUN
bracis-33576	150	21	like	like	ADP
bracis-33576	150	22	age	age	NOUN
bracis-33576	150	23	and	and	CCONJ
bracis-33576	150	24	job	job	NOUN
bracis-33576	150	25	,	,	PUNCT
bracis-33576	150	26	along	along	ADP
bracis-33576	150	27	with	with	ADP
bracis-33576	150	28	the	the	DET
bracis-33576	150	29	response	response	NOUN
bracis-33576	150	30	to	to	ADP
bracis-33576	150	31	the	the	DET
bracis-33576	150	32	campaign	campaign	NOUN
bracis-33576	150	33	 	 	SPACE
bracis-33576	151	1	[	[	X
bracis-33576	151	2	10	10	NUM
bracis-33576	151	3	]	]	PUNCT
bracis-33576	151	4	.	.	PUNCT
bracis-33576	152	1	e	e	X
bracis-33576	152	2	-	-	ADJ
bracis-33576	152	3	commerce	commerce	NOUN
bracis-33576	152	4	public	public	NOUN
bracis-33576	152	5	dataset	dataset	VERB
bracis-33576	152	6	by	by	ADP
bracis-33576	152	7	olist	olist	ADJ
bracis-33576	152	8	:	:	PUNCT
bracis-33576	152	9	a	a	DET
bracis-33576	152	10	brazilian	brazilian	ADJ
bracis-33576	152	11	e	e	NOUN
bracis-33576	152	12	-	-	NOUN
bracis-33576	152	13	commerce	commerce	NOUN
bracis-33576	152	14	dataset	dataset	NOUN
bracis-33576	152	15	with	with	ADP
bracis-33576	152	16	over	over	ADP
bracis-33576	152	17	100,000	100,000	NUM
bracis-33576	152	18	orders	order	NOUN
bracis-33576	152	19	from	from	ADP
bracis-33576	152	20	2016	2016	NUM
bracis-33576	152	21	to	to	ADP
bracis-33576	152	22	2018	2018	NUM
bracis-33576	152	23	across	across	ADP
bracis-33576	152	24	multiple	multiple	ADJ
bracis-33576	152	25	marketplaces	marketplace	NOUN
bracis-33576	152	26	 	 	SPACE
bracis-33576	153	1	[	[	X
bracis-33576	153	2	11	11	NUM
bracis-33576	153	3	]	]	PUNCT
bracis-33576	153	4	.	.	PUNCT
bracis-33576	154	1	it	it	PRON
bracis-33576	154	2	includes	include	VERB
bracis-33576	154	3	72,794	72,794	NUM
bracis-33576	154	4	training	training	NOUN
bracis-33576	154	5	and	and	CCONJ
bracis-33576	154	6	18,199	18,199	NUM
bracis-33576	154	7	testing	testing	NOUN
bracis-33576	154	8	samples	sample	NOUN
bracis-33576	154	9	.	.	PUNCT
bracis-33576	155	1	the	the	DET
bracis-33576	155	2	recency	recency	NOUN
bracis-33576	155	3	,	,	PUNCT
bracis-33576	155	4	frequency	frequency	NOUN
bracis-33576	155	5	,	,	PUNCT
bracis-33576	155	6	monetary	monetary	ADJ
bracis-33576	155	7	(	(	PUNCT
bracis-33576	155	8	rfm	rfm	PROPN
bracis-33576	155	9	)	)	PUNCT
bracis-33576	155	10	model	model	NOUN
bracis-33576	155	11	was	be	AUX
bracis-33576	155	12	used	use	VERB
bracis-33576	155	13	for	for	ADP
bracis-33576	155	14	customer	customer	NOUN
bracis-33576	155	15	segmentation	segmentation	NOUN
bracis-33576	155	16	,	,	PUNCT
bracis-33576	155	17	as	as	SCONJ
bracis-33576	155	18	described	describe	VERB
bracis-33576	155	19	in	in	ADP
bracis-33576	155	20	 	 	SPACE
bracis-33576	155	21	[	[	X
bracis-33576	155	22	19	19	NUM
bracis-33576	155	23	]	]	PUNCT
bracis-33576	155	24	.	.	PUNCT
bracis-33576	156	1	yelp	yelp	NOUN
bracis-33576	156	2	:	:	PUNCT
bracis-33576	156	3	comprises	comprise	VERB
bracis-33576	156	4	reviews	review	NOUN
bracis-33576	156	5	from	from	ADP
bracis-33576	156	6	yelp	yelp	NOUN
bracis-33576	156	7	businesses	business	NOUN
bracis-33576	156	8	,	,	PUNCT
bracis-33576	156	9	including	include	VERB
bracis-33576	156	10	text	text	NOUN
bracis-33576	156	11	reviews	review	NOUN
bracis-33576	156	12	,	,	PUNCT
bracis-33576	156	13	star	star	NOUN
bracis-33576	156	14	ratings	rating	NOUN
bracis-33576	156	15	,	,	PUNCT
bracis-33576	156	16	and	and	CCONJ
bracis-33576	156	17	business	business	NOUN
bracis-33576	156	18	attributes	attribute	NOUN
bracis-33576	156	19	,	,	PUNCT
bracis-33576	156	20	offering	offer	VERB
bracis-33576	156	21	a	a	DET
bracis-33576	156	22	rich	rich	ADJ
bracis-33576	156	23	resource	resource	NOUN
bracis-33576	156	24	for	for	ADP
bracis-33576	156	25	sentiment	sentiment	NOUN
bracis-33576	156	26	analysis	analysis	NOUN
bracis-33576	156	27	and	and	CCONJ
bracis-33576	156	28	review	review	VERB
bracis-33576	156	29	classification	classification	NOUN
bracis-33576	156	30	tasks	task	NOUN
bracis-33576	156	31	 	 	SPACE
bracis-33576	157	1	[	[	X
bracis-33576	157	2	4	4	NUM
bracis-33576	157	3	]	]	PUNCT
bracis-33576	157	4	.	.	PUNCT
bracis-33576	158	1	petfinder.my	petfinder.my	PROPN
bracis-33576	158	2	:	:	PUNCT
bracis-33576	158	3	features	feature	VERB
bracis-33576	158	4	adoption	adoption	NOUN
bracis-33576	158	5	records	record	NOUN
bracis-33576	158	6	from	from	ADP
bracis-33576	158	7	the	the	DET
bracis-33576	158	8	petfinder.my	petfinder.my	PROPN
bracis-33576	158	9	website	website	NOUN
bracis-33576	158	10	,	,	PUNCT
bracis-33576	158	11	encompassing	encompass	VERB
bracis-33576	158	12	various	various	ADJ
bracis-33576	158	13	pet	pet	ADJ
bracis-33576	158	14	attributes	attribute	NOUN
bracis-33576	158	15	,	,	PUNCT
bracis-33576	158	16	descriptions	description	NOUN
bracis-33576	158	17	,	,	PUNCT
bracis-33576	158	18	and	and	CCONJ
bracis-33576	158	19	adoption	adoption	NOUN
bracis-33576	158	20	status	status	NOUN
bracis-33576	158	21	,	,	PUNCT
bracis-33576	158	22	valuable	valuable	ADJ
bracis-33576	158	23	for	for	ADP
bracis-33576	158	24	text	text	NOUN
bracis-33576	158	25	classification	classification	NOUN
bracis-33576	158	26	and	and	CCONJ
bracis-33576	158	27	clustering	clustering	NOUN
bracis-33576	158	28	related	relate	VERB
bracis-33576	158	29	to	to	ADP
bracis-33576	158	30	animal	animal	NOUN
bracis-33576	158	31	welfare	welfare	NOUN
bracis-33576	158	32	 	 	SPACE
bracis-33576	159	1	[	[	X
bracis-33576	159	2	5	5	NUM
bracis-33576	159	3	]	]	PUNCT
bracis-33576	159	4	.	.	PUNCT
bracis-33576	160	1	women	woman	NOUN
bracis-33576	160	2	clothing	clothing	NOUN
bracis-33576	160	3	reviews	review	NOUN
bracis-33576	160	4	:	:	PUNCT
bracis-33576	160	5	contains	contain	VERB
bracis-33576	160	6	reviews	review	NOUN
bracis-33576	160	7	of	of	ADP
bracis-33576	160	8	women	woman	NOUN
bracis-33576	160	9	’s	’s	PART
bracis-33576	160	10	clothing	clothing	NOUN
bracis-33576	160	11	,	,	PUNCT
bracis-33576	160	12	with	with	ADP
bracis-33576	160	13	each	each	DET
bracis-33576	160	14	review	review	NOUN
bracis-33576	160	15	detailing	detail	VERB
bracis-33576	160	16	text	text	NOUN
bracis-33576	160	17	feedback	feedback	NOUN
bracis-33576	160	18	,	,	PUNCT
bracis-33576	160	19	ratings	rating	NOUN
bracis-33576	160	20	,	,	PUNCT
bracis-33576	160	21	and	and	CCONJ
bracis-33576	160	22	customer	customer	NOUN
bracis-33576	160	23	information	information	NOUN
bracis-33576	160	24	,	,	PUNCT
bracis-33576	160	25	suitable	suitable	ADJ
bracis-33576	160	26	for	for	ADP
bracis-33576	160	27	sentiment	sentiment	NOUN
bracis-33576	160	28	analysis	analysis	NOUN
bracis-33576	160	29	and	and	CCONJ
bracis-33576	160	30	recommendation	recommendation	NOUN
bracis-33576	160	31	system	system	NOUN
bracis-33576	160	32	research	research	NOUN
bracis-33576	160	33	 	 	SPACE
bracis-33576	161	1	[	[	X
bracis-33576	161	2	3	3	NUM
bracis-33576	161	3	]	]	PUNCT
bracis-33576	161	4	.	.	PUNCT
bracis-33576	162	1	each	each	DET
bracis-33576	162	2	unlabeled	unlabele	VERB
bracis-33576	162	3	dataset	dataset	NOUN
bracis-33576	162	4	was	be	AUX
bracis-33576	162	5	processed	process	VERB
bracis-33576	162	6	through	through	ADP
bracis-33576	162	7	the	the	DET
bracis-33576	162	8	proposed	propose	VERB
bracis-33576	162	9	pipeline	pipeline	NOUN
bracis-33576	162	10	,	,	PUNCT
bracis-33576	162	11	which	which	PRON
bracis-33576	162	12	involves	involve	VERB
bracis-33576	162	13	training	train	VERB
bracis-33576	162	14	a	a	DET
bracis-33576	162	15	pretrained	pretraine	VERB
bracis-33576	162	16	llm	llm	NOUN
bracis-33576	162	17	.	.	PUNCT
bracis-33576	163	1	this	this	DET
bracis-33576	163	2	approach	approach	NOUN
bracis-33576	163	3	ensures	ensure	VERB
bracis-33576	163	4	that	that	SCONJ
bracis-33576	163	5	the	the	DET
bracis-33576	163	6	clustering	cluster	VERB
bracis-33576	163	7	algorithms	algorithm	NOUN
bracis-33576	163	8	can	can	AUX
bracis-33576	163	9	perform	perform	VERB
bracis-33576	163	10	optimally	optimally	ADV
bracis-33576	163	11	across	across	ADP
bracis-33576	163	12	diverse	diverse	ADJ
bracis-33576	163	13	textual	textual	ADJ
bracis-33576	163	14	inputs	input	NOUN
bracis-33576	163	15	,	,	PUNCT
bracis-33576	163	16	enhancing	enhance	VERB
bracis-33576	163	17	their	their	PRON
bracis-33576	163	18	ability	ability	NOUN
bracis-33576	163	19	to	to	PART
bracis-33576	163	20	effectively	effectively	ADV
bracis-33576	163	21	identify	identify	VERB
bracis-33576	163	22	and	and	CCONJ
bracis-33576	163	23	group	group	NOUN
bracis-33576	163	24	related	relate	VERB
bracis-33576	163	25	data	datum	NOUN
bracis-33576	163	26	points	point	NOUN
bracis-33576	163	27	.	.	PUNCT
bracis-33576	164	1	4.2	4.2	NUM
bracis-33576	164	2	clustering	cluster	VERB
bracis-33576	164	3	algorithms	algorithm	NOUN
bracis-33576	164	4	the	the	DET
bracis-33576	164	5	clustering	cluster	VERB
bracis-33576	164	6	algorithms	algorithm	NOUN
bracis-33576	164	7	chosen	choose	VERB
bracis-33576	164	8	are	be	AUX
bracis-33576	164	9	well	well	ADV
bracis-33576	164	10	-	-	PUNCT
bracis-33576	164	11	suited	suit	VERB
bracis-33576	164	12	for	for	ADP
bracis-33576	164	13	handling	handle	VERB
bracis-33576	164	14	complex	complex	ADJ
bracis-33576	164	15	patterns	pattern	NOUN
bracis-33576	164	16	in	in	ADP
bracis-33576	164	17	structured	structured	ADJ
bracis-33576	164	18	and	and	CCONJ
bracis-33576	164	19	textual	textual	ADJ
bracis-33576	164	20	data	datum	NOUN
bracis-33576	164	21	,	,	PUNCT
bracis-33576	164	22	ensuring	ensure	VERB
bracis-33576	164	23	efficient	efficient	ADJ
bracis-33576	164	24	categorization	categorization	NOUN
bracis-33576	164	25	.	.	PUNCT
bracis-33576	165	1	we	we	PRON
bracis-33576	165	2	used	use	VERB
bracis-33576	165	3	the	the	DET
bracis-33576	165	4	k	k	NOUN
bracis-33576	165	5	-	-	PUNCT
bracis-33576	165	6	means	means	NOUN
bracis-33576	165	7	algorithm	algorithm	NOUN
bracis-33576	165	8	for	for	ADP
bracis-33576	165	9	its	its	PRON
bracis-33576	165	10	simplicity	simplicity	NOUN
bracis-33576	165	11	and	and	CCONJ
bracis-33576	165	12	effectiveness	effectiveness	NOUN
bracis-33576	165	13	with	with	ADP
bracis-33576	165	14	large	large	ADJ
bracis-33576	165	15	datasets	dataset	NOUN
bracis-33576	165	16	and	and	CCONJ
bracis-33576	165	17	k	k	NOUN
bracis-33576	165	18	-	-	NOUN
bracis-33576	165	19	means++	means++	NOUN
bracis-33576	165	20	for	for	ADP
bracis-33576	165	21	its	its	PRON
bracis-33576	165	22	strategic	strategic	ADJ
bracis-33576	165	23	centroid	centroid	NOUN
bracis-33576	165	24	initialization	initialization	NOUN
bracis-33576	165	25	to	to	PART
bracis-33576	165	26	enhance	enhance	VERB
bracis-33576	165	27	clustering	clustering	ADJ
bracis-33576	165	28	efficiency	efficiency	NOUN
bracis-33576	165	29	and	and	CCONJ
bracis-33576	165	30	quality	quality	NOUN
bracis-33576	165	31	 	 	SPACE
bracis-33576	165	32	[	[	X
bracis-33576	165	33	12	12	NUM
bracis-33576	165	34	]	]	PUNCT
bracis-33576	165	35	.	.	PUNCT
bracis-33576	166	1	unlike	unlike	ADP
bracis-33576	166	2	k	k	NOUN
bracis-33576	166	3	-	-	PUNCT
bracis-33576	166	4	means	mean	NOUN
bracis-33576	166	5	,	,	PUNCT
bracis-33576	166	6	which	which	PRON
bracis-33576	166	7	assigns	assign	VERB
bracis-33576	166	8	each	each	DET
bracis-33576	166	9	data	data	NOUN
bracis-33576	166	10	point	point	NOUN
bracis-33576	166	11	to	to	ADP
bracis-33576	166	12	a	a	DET
bracis-33576	166	13	single	single	ADJ
bracis-33576	166	14	cluster	cluster	NOUN
bracis-33576	166	15	,	,	PUNCT
bracis-33576	166	16	fuzzy	fuzzy	ADJ
bracis-33576	166	17	c	c	NOUN
bracis-33576	166	18	-	-	PUNCT
bracis-33576	166	19	means	means	NOUN
bracis-33576	166	20	(	(	PUNCT
bracis-33576	166	21	fuzzycm	fuzzycm	PROPN
bracis-33576	166	22	)	)	PUNCT
bracis-33576	166	23	employs	employ	VERB
bracis-33576	166	24	a	a	DET
bracis-33576	166	25	probabilistic	probabilistic	ADJ
bracis-33576	166	26	membership	membership	NOUN
bracis-33576	166	27	approach	approach	NOUN
bracis-33576	166	28	,	,	PUNCT
bracis-33576	166	29	effectively	effectively	ADV
bracis-33576	166	30	capturing	capture	VERB
bracis-33576	166	31	the	the	DET
bracis-33576	166	32	nuances	nuance	NOUN
bracis-33576	166	33	and	and	CCONJ
bracis-33576	166	34	polysemy	polysemy	NOUN
bracis-33576	166	35	typical	typical	ADJ
bracis-33576	166	36	of	of	ADP
bracis-33576	166	37	textual	textual	ADJ
bracis-33576	166	38	data	datum	NOUN
bracis-33576	166	39	.	.	PUNCT
bracis-33576	167	1	we	we	PRON
bracis-33576	167	2	used	use	VERB
bracis-33576	167	3	agglomerative	agglomerative	ADJ
bracis-33576	167	4	hierarchical	hierarchical	ADJ
bracis-33576	167	5	clustering	clustering	NOUN
bracis-33576	167	6	(	(	PUNCT
bracis-33576	167	7	ahc	ahc	NOUN
bracis-33576	167	8	)	)	PUNCT
bracis-33576	167	9	to	to	PART
bracis-33576	167	10	uncover	uncover	VERB
bracis-33576	167	11	hierarchical	hierarchical	ADJ
bracis-33576	167	12	structures	structure	NOUN
bracis-33576	167	13	and	and	CCONJ
bracis-33576	167	14	spectral	spectral	ADJ
bracis-33576	167	15	clustering	clustering	NOUN
bracis-33576	167	16	for	for	ADP
bracis-33576	167	17	its	its	PRON
bracis-33576	167	18	proficiency	proficiency	NOUN
bracis-33576	167	19	in	in	ADP
bracis-33576	167	20	recognizing	recognize	VERB
bracis-33576	167	21	clusters	cluster	NOUN
bracis-33576	167	22	based	base	VERB
bracis-33576	167	23	on	on	ADP
bracis-33576	167	24	the	the	DET
bracis-33576	167	25	data	data	NOUN
bracis-33576	167	26	’s	’s	PART
bracis-33576	167	27	graph	graph	NOUN
bracis-33576	167	28	structure	structure	NOUN
bracis-33576	167	29	,	,	PUNCT
bracis-33576	167	30	effectively	effectively	ADV
bracis-33576	167	31	identifying	identify	VERB
bracis-33576	167	32	non	non	ADJ
bracis-33576	167	33	-	-	ADJ
bracis-33576	167	34	convex	convex	ADJ
bracis-33576	167	35	shapes	shape	NOUN
bracis-33576	167	36	.	.	PUNCT
bracis-33576	168	1	for	for	ADP
bracis-33576	168	2	k	k	NOUN
bracis-33576	168	3	-	-	PUNCT
bracis-33576	168	4	means	mean	NOUN
bracis-33576	168	5	,	,	PUNCT
bracis-33576	168	6	the	the	DET
bracis-33576	168	7	parameters	parameter	NOUN
bracis-33576	168	8	were	be	AUX
bracis-33576	168	9	:	:	PUNCT
bracis-33576	168	10	initialization	initialization	NOUN
bracis-33576	168	11	method	method	NOUN
bracis-33576	168	12	set	set	VERB
bracis-33576	168	13	to	to	ADP
bracis-33576	168	14	random	random	ADJ
bracis-33576	168	15	,	,	PUNCT
bracis-33576	168	16	number	number	NOUN
bracis-33576	168	17	of	of	ADP
bracis-33576	168	18	initializations	initialization	NOUN
bracis-33576	168	19	(	(	PUNCT
bracis-33576	168	20	ninit	ninit	NOUN
bracis-33576	168	21	)	)	PUNCT
bracis-33576	168	22	set	set	VERB
bracis-33576	168	23	to	to	ADP
bracis-33576	168	24	10	10	NUM
bracis-33576	168	25	,	,	PUNCT
bracis-33576	168	26	and	and	CCONJ
bracis-33576	168	27	the	the	DET
bracis-33576	168	28	random	random	ADJ
bracis-33576	168	29	seed	seed	NOUN
bracis-33576	168	30	set	set	VERB
bracis-33576	168	31	to	to	ADP
bracis-33576	168	32	0	0	NUM
bracis-33576	168	33	.	.	PUNCT
bracis-33576	169	1	the	the	DET
bracis-33576	169	2	k	k	ADJ
bracis-33576	169	3	-	-	ADJ
bracis-33576	169	4	means++	means++	ADJ
bracis-33576	169	5	algorithm	algorithm	NOUN
bracis-33576	169	6	utilized	utilize	VERB
bracis-33576	169	7	k	k	X
bracis-33576	169	8	-	-	NOUN
bracis-33576	169	9	means++	means++	PROPN
bracis-33576	169	10	for	for	ADP
bracis-33576	169	11	initialization	initialization	NOUN
bracis-33576	169	12	,	,	PUNCT
bracis-33576	169	13	ninit	ninit	PROPN
bracis-33576	169	14	was	be	AUX
bracis-33576	169	15	1	1	NUM
bracis-33576	169	16	,	,	PUNCT
bracis-33576	169	17	and	and	CCONJ
bracis-33576	169	18	the	the	DET
bracis-33576	169	19	seed	seed	NOUN
bracis-33576	169	20	was	be	AUX
bracis-33576	169	21	0	0	NUM
bracis-33576	169	22	.	.	PUNCT
bracis-33576	170	1	agglomerative	agglomerative	ADJ
bracis-33576	170	2	hierarchical	hierarchical	ADJ
bracis-33576	170	3	clustering	clustering	NOUN
bracis-33576	170	4	(	(	PUNCT
bracis-33576	170	5	ahc	ahc	NOUN
bracis-33576	170	6	)	)	PUNCT
bracis-33576	170	7	employed	employ	VERB
bracis-33576	170	8	the	the	DET
bracis-33576	170	9	euclidean	euclidean	ADJ
bracis-33576	170	10	metric	metric	NOUN
bracis-33576	170	11	with	with	ADP
bracis-33576	170	12	ward	ward	NOUN
bracis-33576	170	13	linkage	linkage	NOUN
bracis-33576	170	14	.	.	PUNCT
bracis-33576	171	1	for	for	ADP
bracis-33576	171	2	fuzzy	fuzzy	ADJ
bracis-33576	171	3	c	c	NOUN
bracis-33576	171	4	-	-	PUNCT
bracis-33576	171	5	means	means	NOUN
bracis-33576	171	6	(	(	PUNCT
bracis-33576	171	7	fuzzycm	fuzzycm	PROPN
bracis-33576	171	8	)	)	PUNCT
bracis-33576	171	9	,	,	PUNCT
bracis-33576	171	10	no	no	DET
bracis-33576	171	11	initialization	initialization	NOUN
bracis-33576	171	12	method	method	NOUN
bracis-33576	171	13	was	be	AUX
bracis-33576	171	14	specified	specify	VERB
bracis-33576	171	15	,	,	PUNCT
bracis-33576	171	16	the	the	DET
bracis-33576	171	17	fuzziness	fuzziness	NOUN
bracis-33576	171	18	parameter	parameter	NOUN
bracis-33576	171	19	(	(	PUNCT
bracis-33576	171	20	m	m	NOUN
bracis-33576	171	21	)	)	PUNCT
bracis-33576	171	22	was	be	AUX
bracis-33576	171	23	2	2	NUM
bracis-33576	171	24	,	,	PUNCT
bracis-33576	171	25	error	error	NOUN
bracis-33576	171	26	tolerance	tolerance	NOUN
bracis-33576	171	27	was	be	AUX
bracis-33576	171	28	set	set	VERB
bracis-33576	171	29	to	to	ADP
bracis-33576	171	30	0.005	0.005	NUM
bracis-33576	171	31	,	,	PUNCT
bracis-33576	171	32	and	and	CCONJ
bracis-33576	171	33	the	the	DET
bracis-33576	171	34	maximum	maximum	ADJ
bracis-33576	171	35	number	number	NOUN
bracis-33576	171	36	of	of	ADP
bracis-33576	171	37	iterations	iteration	NOUN
bracis-33576	171	38	(	(	PUNCT
bracis-33576	171	39	maxiter	maxiter	NOUN
bracis-33576	171	40	)	)	PUNCT
bracis-33576	171	41	was	be	AUX
bracis-33576	171	42	1000	1000	NUM
bracis-33576	171	43	.	.	PUNCT
bracis-33576	172	1	spectral	spectral	ADJ
bracis-33576	172	2	clustering	clustering	NOUN
bracis-33576	172	3	used	use	VERB
bracis-33576	172	4	’	'	PUNCT
bracis-33576	172	5	discretize	discretize	NOUN
bracis-33576	172	6	’	'	PUNCT
bracis-33576	172	7	for	for	ADP
bracis-33576	172	8	label	label	NOUN
bracis-33576	172	9	assignment	assignment	NOUN
bracis-33576	172	10	and	and	CCONJ
bracis-33576	172	11	a	a	DET
bracis-33576	172	12	random	random	ADJ
bracis-33576	172	13	seed	seed	NOUN
bracis-33576	172	14	of	of	ADP
bracis-33576	172	15	10	10	NUM
bracis-33576	172	16	.	.	PUNCT
bracis-33576	173	1	implementations	implementation	NOUN
bracis-33576	173	2	for	for	ADP
bracis-33576	173	3	these	these	DET
bracis-33576	173	4	algorithms	algorithm	NOUN
bracis-33576	173	5	were	be	AUX
bracis-33576	173	6	sourced	source	VERB
bracis-33576	173	7	from	from	ADP
bracis-33576	173	8	the	the	DET
bracis-33576	173	9	scikit	scikit	NOUN
bracis-33576	173	10	-	-	PUNCT
bracis-33576	173	11	learn	learn	VERB
bracis-33576	173	12	library	library	NOUN
bracis-33576	173	13	[	[	X
bracis-33576	173	14	13	13	NUM
bracis-33576	173	15	]	]	PUNCT
bracis-33576	173	16	,	,	PUNCT
bracis-33576	173	17	except	except	SCONJ
bracis-33576	173	18	for	for	ADP
bracis-33576	173	19	fuzzycm	fuzzycm	PROPN
bracis-33576	173	20	,	,	PUNCT
bracis-33576	173	21	which	which	PRON
bracis-33576	173	22	used	use	VERB
bracis-33576	173	23	the	the	DET
bracis-33576	173	24	scikit	scikit	NOUN
bracis-33576	173	25	-	-	PUNCT
bracis-33576	173	26	fuzzy	fuzzy	ADJ
bracis-33576	173	27	package	package	NOUN
bracis-33576	173	28	 	 	SPACE
bracis-33576	174	1	[	[	X
bracis-33576	174	2	22	22	NUM
bracis-33576	174	3	]	]	PUNCT
bracis-33576	174	4	.	.	PUNCT
bracis-33576	175	1	for	for	ADP
bracis-33576	175	2	k	k	NOUN
bracis-33576	175	3	-	-	PUNCT
bracis-33576	175	4	means	mean	NOUN
bracis-33576	175	5	and	and	CCONJ
bracis-33576	175	6	k	k	NOUN
bracis-33576	175	7	-	-	PUNCT
bracis-33576	175	8	means++	means++	PROPN
bracis-33576	175	9	,	,	PUNCT
bracis-33576	175	10	init	init	PROPN
bracis-33576	175	11	specifies	specify	VERB
bracis-33576	175	12	the	the	DET
bracis-33576	175	13	initial	initial	ADJ
bracis-33576	175	14	cluster	cluster	NOUN
bracis-33576	175	15	centroid	centroid	NOUN
bracis-33576	175	16	method	method	NOUN
bracis-33576	175	17	,	,	PUNCT
bracis-33576	175	18	ninit	ninit	PROPN
bracis-33576	175	19	indicates	indicate	VERB
bracis-33576	175	20	the	the	DET
bracis-33576	175	21	number	number	NOUN
bracis-33576	175	22	of	of	ADP
bracis-33576	175	23	algorithm	algorithm	NOUN
bracis-33576	175	24	runs	run	VERB
bracis-33576	175	25	with	with	ADP
bracis-33576	175	26	different	different	ADJ
bracis-33576	175	27	seeds	seed	NOUN
bracis-33576	175	28	,	,	PUNCT
bracis-33576	175	29	and	and	CCONJ
bracis-33576	175	30	seed	seed	NOUN
bracis-33576	175	31	sets	set	VERB
bracis-33576	175	32	the	the	DET
bracis-33576	175	33	random	random	ADJ
bracis-33576	175	34	number	number	NOUN
bracis-33576	175	35	for	for	ADP
bracis-33576	175	36	centroid	centroid	NOUN
bracis-33576	175	37	initialization	initialization	NOUN
bracis-33576	175	38	.	.	PUNCT
bracis-33576	176	1	in	in	ADP
bracis-33576	176	2	ahc	ahc	NOUN
bracis-33576	176	3	,	,	PUNCT
bracis-33576	176	4	metric	metric	ADJ
bracis-33576	176	5	is	be	AUX
bracis-33576	176	6	the	the	DET
bracis-33576	176	7	metric	metric	ADJ
bracis-33576	176	8	used	use	VERB
bracis-33576	176	9	for	for	ADP
bracis-33576	176	10	linkage	linkage	NOUN
bracis-33576	176	11	computation	computation	NOUN
bracis-33576	176	12	,	,	PUNCT
bracis-33576	176	13	and	and	CCONJ
bracis-33576	176	14	linkage	linkage	NOUN
bracis-33576	176	15	is	be	AUX
bracis-33576	176	16	the	the	DET
bracis-33576	176	17	criterion	criterion	NOUN
bracis-33576	176	18	measuring	measure	VERB
bracis-33576	176	19	the	the	DET
bracis-33576	176	20	distance	distance	NOUN
bracis-33576	176	21	between	between	ADP
bracis-33576	176	22	observation	observation	NOUN
bracis-33576	176	23	sets	set	NOUN
bracis-33576	176	24	.	.	PUNCT
bracis-33576	177	1	we	we	PRON
bracis-33576	177	2	used	use	VERB
bracis-33576	177	3	euclidean	euclidean	ADJ
bracis-33576	177	4	distance	distance	NOUN
bracis-33576	177	5	to	to	PART
bracis-33576	177	6	measure	measure	NOUN
bracis-33576	177	7	point	point	NOUN
bracis-33576	177	8	similarities	similarity	NOUN
bracis-33576	177	9	and	and	CCONJ
bracis-33576	177	10	a	a	DET
bracis-33576	177	11	nearest	near	ADJ
bracis-33576	177	12	centroid	centroid	NOUN
bracis-33576	177	13	approach	approach	NOUN
bracis-33576	177	14	to	to	ADP
bracis-33576	177	15	associate	associate	ADJ
bracis-33576	177	16	clusters	cluster	NOUN
bracis-33576	177	17	.	.	PUNCT
bracis-33576	178	1	for	for	ADP
bracis-33576	178	2	fuzzycm	fuzzycm	PROPN
bracis-33576	178	3	,	,	PUNCT
bracis-33576	178	4	init	init	PROPN
bracis-33576	178	5	is	be	AUX
bracis-33576	178	6	the	the	DET
bracis-33576	178	7	initial	initial	ADJ
bracis-33576	178	8	fuzzy	fuzzy	ADJ
bracis-33576	178	9	c	c	NOUN
bracis-33576	178	10	-	-	PUNCT
bracis-33576	178	11	partitioned	partition	VERB
bracis-33576	178	12	matrix	matrix	NOUN
bracis-33576	178	13	(	(	PUNCT
bracis-33576	178	14	random	random	ADJ
bracis-33576	178	15	if	if	SCONJ
bracis-33576	178	16	none	none	NOUN
bracis-33576	178	17	)	)	PUNCT
bracis-33576	178	18	,	,	PUNCT
bracis-33576	178	19	m	m	VERB
bracis-33576	178	20	is	be	AUX
bracis-33576	178	21	the	the	DET
bracis-33576	178	22	fuzziness	fuzziness	NOUN
bracis-33576	178	23	degree	degree	NOUN
bracis-33576	178	24	,	,	PUNCT
bracis-33576	178	25	error	error	NOUN
bracis-33576	178	26	is	be	AUX
bracis-33576	178	27	the	the	DET
bracis-33576	178	28	stopping	stopping	NOUN
bracis-33576	178	29	criterion	criterion	NOUN
bracis-33576	178	30	,	,	PUNCT
bracis-33576	178	31	and	and	CCONJ
bracis-33576	178	32	maxiter	maxiter	NOUN
bracis-33576	178	33	is	be	AUX
bracis-33576	178	34	the	the	DET
bracis-33576	178	35	iteration	iteration	NOUN
bracis-33576	178	36	limit	limit	NOUN
bracis-33576	178	37	.	.	PUNCT
bracis-33576	179	1	in	in	ADP
bracis-33576	179	2	spectral	spectral	ADJ
bracis-33576	179	3	clustering	clustering	NOUN
bracis-33576	179	4	,	,	PUNCT
bracis-33576	179	5	assign_labels	assign_label	NOUN
bracis-33576	179	6	specifies	specify	VERB
bracis-33576	179	7	the	the	DET
bracis-33576	179	8	labeling	labeling	NOUN
bracis-33576	179	9	strategy	strategy	NOUN
bracis-33576	179	10	in	in	ADP
bracis-33576	179	11	the	the	DET
bracis-33576	179	12	embedding	embed	VERB
bracis-33576	179	13	space	space	NOUN
bracis-33576	179	14	,	,	PUNCT
bracis-33576	179	15	and	and	CCONJ
bracis-33576	179	16	seed	seed	NOUN
bracis-33576	179	17	is	be	AUX
bracis-33576	179	18	the	the	DET
bracis-33576	179	19	pseudorandom	pseudorandom	ADJ
bracis-33576	179	20	number	number	NOUN
bracis-33576	179	21	for	for	ADP
bracis-33576	179	22	initializing	initialize	VERB
bracis-33576	179	23	the	the	DET
bracis-33576	179	24	eigenvector	eigenvector	NOUN
bracis-33576	179	25	decomposition	decomposition	NOUN
bracis-33576	179	26	.	.	PUNCT
bracis-33576	180	1	for	for	ADP
bracis-33576	180	2	all	all	DET
bracis-33576	180	3	datasets	dataset	NOUN
bracis-33576	180	4	,	,	PUNCT
bracis-33576	180	5	the	the	DET
bracis-33576	180	6	number	number	NOUN
bracis-33576	180	7	of	of	ADP
bracis-33576	180	8	clusters	cluster	NOUN
bracis-33576	180	9	(	(	PUNCT
bracis-33576	180	10	\	\	PROPN
bracis-33576	180	11	(	(	PUNCT
bracis-33576	180	12	k	k	PROPN
bracis-33576	180	13	\	\	PROPN
bracis-33576	180	14	)	)	PUNCT
bracis-33576	180	15	)	)	PUNCT
bracis-33576	180	16	was	be	AUX
bracis-33576	180	17	determined	determine	VERB
bracis-33576	180	18	using	use	VERB
bracis-33576	180	19	the	the	DET
bracis-33576	180	20	silhouette	silhouette	NOUN
bracis-33576	180	21	score	score	NOUN
bracis-33576	180	22	to	to	PART
bracis-33576	180	23	optimize	optimize	VERB
bracis-33576	180	24	cluster	cluster	NOUN
bracis-33576	180	25	cohesion	cohesion	NOUN
bracis-33576	180	26	and	and	CCONJ
bracis-33576	180	27	separation	separation	NOUN
bracis-33576	180	28	.	.	PUNCT
bracis-33576	181	1	4.3	4.3	NUM
bracis-33576	181	2	approaches	approach	NOUN
bracis-33576	181	3	(	(	PUNCT
bracis-33576	181	4	baselines	baseline	NOUN
bracis-33576	181	5	)	)	PUNCT
bracis-33576	181	6	we	we	PRON
bracis-33576	181	7	utilized	utilize	VERB
bracis-33576	181	8	various	various	ADJ
bracis-33576	181	9	embedding	embed	VERB
bracis-33576	181	10	techniques	technique	NOUN
bracis-33576	181	11	from	from	ADP
bracis-33576	181	12	state	state	NOUN
bracis-33576	181	13	-	-	PUNCT
bracis-33576	181	14	of	of	ADP
bracis-33576	181	15	-	-	PUNCT
bracis-33576	181	16	the	the	DET
bracis-33576	181	17	-	-	PUNCT
bracis-33576	181	18	art	art	NOUN
bracis-33576	181	19	llms	llm	NOUN
bracis-33576	181	20	,	,	PUNCT
bracis-33576	181	21	including	include	VERB
bracis-33576	181	22	openai	openai	PROPN
bracis-33576	181	23	,	,	PUNCT
bracis-33576	181	24	falcon	falcon	NOUN
bracis-33576	181	25	,	,	PUNCT
bracis-33576	181	26	llama	llama	NOUN
bracis-33576	181	27	2	2	NUM
bracis-33576	181	28	,	,	PUNCT
bracis-33576	181	29	gpt-2	gpt-2	X
bracis-33576	181	30	medium	medium	NOUN
bracis-33576	181	31	,	,	PUNCT
bracis-33576	181	32	and	and	CCONJ
bracis-33576	181	33	an	an	DET
bracis-33576	181	34	mpnet	mpnet	NOUN
bracis-33576	181	35	-	-	PUNCT
bracis-33576	181	36	based	base	VERB
bracis-33576	181	37	model	model	NOUN
bracis-33576	181	38	,	,	PUNCT
bracis-33576	181	39	each	each	PRON
bracis-33576	181	40	enhancing	enhance	VERB
bracis-33576	181	41	text	text	NOUN
bracis-33576	181	42	representation	representation	NOUN
bracis-33576	181	43	by	by	ADP
bracis-33576	181	44	capturing	capture	VERB
bracis-33576	181	45	contextual	contextual	ADJ
bracis-33576	181	46	nuances	nuance	NOUN
bracis-33576	181	47	.	.	PUNCT
bracis-33576	182	1	for	for	ADP
bracis-33576	182	2	the	the	DET
bracis-33576	182	3	mpnet	mpnet	NOUN
bracis-33576	182	4	-	-	PUNCT
bracis-33576	182	5	based	base	VERB
bracis-33576	182	6	model	model	NOUN
bracis-33576	182	7	,	,	PUNCT
bracis-33576	182	8	we	we	PRON
bracis-33576	182	9	used	use	VERB
bracis-33576	182	10	sentence	sentence	NOUN
bracis-33576	182	11	-	-	PUNCT
bracis-33576	182	12	transformers	transformer	NOUN
bracis-33576	182	13	/	/	SYM
bracis-33576	182	14	all	all	DET
bracis-33576	182	15	-	-	PUNCT
bracis-33576	182	16	mpnet	mpnet	NOUN
bracis-33576	182	17	-	-	PUNCT
bracis-33576	182	18	base	base	NOUN
bracis-33576	182	19	-	-	PUNCT
bracis-33576	182	20	v2	v2	NOUN
bracis-33576	182	21	(	(	PUNCT
bracis-33576	182	22	mpnet	mpnet	NOUN
bracis-33576	182	23	-	-	PUNCT
bracis-33576	182	24	v2	v2	NOUN
bracis-33576	182	25	)	)	PUNCT
bracis-33576	182	26	 	 	SPACE
bracis-33576	183	1	[	[	X
bracis-33576	183	2	17	17	NUM
bracis-33576	183	3	]	]	PUNCT
bracis-33576	183	4	.	.	PUNCT
bracis-33576	184	1	for	for	ADP
bracis-33576	184	2	the	the	DET
bracis-33576	184	3	openai	openai	PROPN
bracis-33576	184	4	model	model	NOUN
bracis-33576	184	5	,	,	PUNCT
bracis-33576	184	6	we	we	PRON
bracis-33576	184	7	utilized	utilize	VERB
bracis-33576	184	8	text	text	NOUN
bracis-33576	184	9	-	-	PUNCT
bracis-33576	184	10	embedding-3	embedding-3	NUM
bracis-33576	184	11	-	-	PUNCT
bracis-33576	184	12	large	large	ADJ
bracis-33576	184	13	,	,	PUNCT
bracis-33576	184	14	and	and	CCONJ
bracis-33576	184	15	for	for	ADP
bracis-33576	184	16	the	the	DET
bracis-33576	184	17	falcon	falcon	PROPN
bracis-33576	184	18	model	model	NOUN
bracis-33576	184	19	,	,	PUNCT
bracis-33576	184	20	we	we	PRON
bracis-33576	184	21	used	use	VERB
bracis-33576	184	22	tiiuae	tiiuae	PROPN
bracis-33576	184	23	/	/	SYM
bracis-33576	184	24	falcon-7b	falcon-7b	NOUN
bracis-33576	184	25	 	 	SPACE
bracis-33576	185	1	[	[	X
bracis-33576	185	2	1	1	NUM
bracis-33576	185	3	]	]	PUNCT
bracis-33576	185	4	.	.	PUNCT
bracis-33576	186	1	additionally	additionally	ADV
bracis-33576	186	2	,	,	PUNCT
bracis-33576	186	3	we	we	PRON
bracis-33576	186	4	employed	employ	VERB
bracis-33576	186	5	llama-2	llama-2	PROPN
bracis-33576	186	6	-	-	PUNCT
bracis-33576	186	7	7b	7b	NOUN
bracis-33576	186	8	-	-	PUNCT
bracis-33576	186	9	chat	chat	NOUN
bracis-33576	186	10	-	-	PUNCT
bracis-33576	186	11	hf	hf	NOUN
bracis-33576	186	12	 	 	SPACE
bracis-33576	187	1	[	[	X
bracis-33576	187	2	20	20	NUM
bracis-33576	187	3	]	]	PUNCT
bracis-33576	187	4	for	for	ADP
bracis-33576	187	5	chat	chat	NOUN
bracis-33576	187	6	applications	application	NOUN
bracis-33576	187	7	and	and	CCONJ
bracis-33576	187	8	gpt2	gpt2	NOUN
bracis-33576	187	9	-	-	PUNCT
bracis-33576	187	10	medium	medium	NOUN
bracis-33576	187	11	 	 	SPACE
bracis-33576	187	12	[	[	X
bracis-33576	187	13	16	16	NUM
bracis-33576	187	14	]	]	PUNCT
bracis-33576	187	15	for	for	ADP
bracis-33576	187	16	faster	fast	ADJ
bracis-33576	187	17	textual	textual	ADJ
bracis-33576	187	18	representations	representation	NOUN
bracis-33576	187	19	.	.	PUNCT
bracis-33576	188	1	we	we	PRON
bracis-33576	188	2	used	use	VERB
bracis-33576	188	3	the	the	DET
bracis-33576	188	4	embeddings	embedding	NOUN
bracis-33576	188	5	from	from	ADP
bracis-33576	188	6	the	the	DET
bracis-33576	188	7	last	last	ADJ
bracis-33576	188	8	layer	layer	NOUN
bracis-33576	188	9	of	of	ADP
bracis-33576	188	10	all	all	DET
bracis-33576	188	11	models	model	NOUN
bracis-33576	188	12	for	for	ADP
bracis-33576	188	13	the	the	DET
bracis-33576	188	14	most	most	ADV
bracis-33576	188	15	contextually	contextually	ADV
bracis-33576	188	16	rich	rich	ADJ
bracis-33576	188	17	text	text	NOUN
bracis-33576	188	18	representations	representation	NOUN
bracis-33576	188	19	.	.	PUNCT
bracis-33576	189	1	we	we	PRON
bracis-33576	189	2	integrated	integrate	VERB
bracis-33576	189	3	an	an	DET
bracis-33576	189	4	additional	additional	ADJ
bracis-33576	189	5	baseline	baseline	NOUN
bracis-33576	189	6	from	from	ADP
bracis-33576	189	7	a	a	DET
bracis-33576	189	8	recent	recent	ADJ
bracis-33576	189	9	study	study	NOUN
bracis-33576	189	10	that	that	PRON
bracis-33576	189	11	explores	explore	VERB
bracis-33576	189	12	customer	customer	NOUN
bracis-33576	189	13	segmentation	segmentation	NOUN
bracis-33576	189	14	using	use	VERB
bracis-33576	189	15	llms	llm	NOUN
bracis-33576	189	16	combined	combine	VERB
bracis-33576	189	17	with	with	ADP
bracis-33576	189	18	dice	dice	NOUN
bracis-33576	189	19	 	 	SPACE
bracis-33576	189	20	[	[	X
bracis-33576	189	21	19	19	NUM
bracis-33576	189	22	]	]	PUNCT
bracis-33576	189	23	.	.	PUNCT
bracis-33576	190	1	they	they	PRON
bracis-33576	190	2	used	use	VERB
bracis-33576	190	3	the	the	DET
bracis-33576	190	4	paraphrase	paraphrase	NOUN
bracis-33576	190	5	-	-	PUNCT
bracis-33576	190	6	multilingual	multilingual	NOUN
bracis-33576	190	7	-	-	PUNCT
bracis-33576	190	8	mpnet	mpnet	NOUN
bracis-33576	190	9	-	-	PUNCT
bracis-33576	190	10	base	base	NOUN
bracis-33576	190	11	-	-	PUNCT
bracis-33576	190	12	v2	v2	NOUN
bracis-33576	190	13	model	model	NOUN
bracis-33576	190	14	from	from	ADP
bracis-33576	190	15	sentence	sentence	NOUN
bracis-33576	190	16	transformers	transformer	NOUN
bracis-33576	190	17	to	to	PART
bracis-33576	190	18	generate	generate	VERB
bracis-33576	190	19	768	768	NUM
bracis-33576	190	20	-	-	PUNCT
bracis-33576	190	21	dimensional	dimensional	ADJ
bracis-33576	190	22	sentence	sentence	NOUN
bracis-33576	190	23	embeddings	embedding	NOUN
bracis-33576	190	24	.	.	PUNCT
bracis-33576	191	1	this	this	DET
bracis-33576	191	2	model	model	NOUN
bracis-33576	191	3	,	,	PUNCT
bracis-33576	191	4	based	base	VERB
bracis-33576	191	5	on	on	ADP
bracis-33576	191	6	the	the	DET
bracis-33576	191	7	mpnet	mpnet	ADJ
bracis-33576	191	8	architecture	architecture	NOUN
bracis-33576	191	9	,	,	PUNCT
bracis-33576	191	10	has	have	VERB
bracis-33576	191	11	about	about	ADV
bracis-33576	191	12	278	278	NUM
bracis-33576	191	13	million	million	NUM
bracis-33576	191	14	parameters	parameter	NOUN
bracis-33576	191	15	and	and	CCONJ
bracis-33576	191	16	is	be	AUX
bracis-33576	191	17	designed	design	VERB
bracis-33576	191	18	for	for	ADP
bracis-33576	191	19	clustering	cluster	VERB
bracis-33576	191	20	and	and	CCONJ
bracis-33576	191	21	semantic	semantic	ADJ
bracis-33576	191	22	search	search	NOUN
bracis-33576	191	23	.	.	PUNCT
bracis-33576	192	1	4.4	4.4	NUM
bracis-33576	192	2	evaluation	evaluation	NOUN
bracis-33576	192	3	metrics	metric	NOUN
bracis-33576	192	4	we	we	PRON
bracis-33576	192	5	use	use	VERB
bracis-33576	192	6	a	a	DET
bracis-33576	192	7	set	set	NOUN
bracis-33576	192	8	of	of	ADP
bracis-33576	192	9	metrics	metric	NOUN
bracis-33576	192	10	to	to	PART
bracis-33576	192	11	evaluate	evaluate	VERB
bracis-33576	192	12	our	our	PRON
bracis-33576	192	13	proposed	propose	VERB
bracis-33576	192	14	framework	framework	NOUN
bracis-33576	192	15	and	and	CCONJ
bracis-33576	192	16	baselines	baseline	NOUN
bracis-33576	192	17	thoroughly	thoroughly	ADV
bracis-33576	192	18	.	.	PUNCT
bracis-33576	193	1	specifically	specifically	ADV
bracis-33576	193	2	,	,	PUNCT
bracis-33576	193	3	we	we	PRON
bracis-33576	193	4	employed	employ	VERB
bracis-33576	193	5	the	the	DET
bracis-33576	193	6	ss	ss	PROPN
bracis-33576	193	7	,	,	PUNCT
bracis-33576	193	8	chi	chi	NOUN
bracis-33576	193	9	,	,	PUNCT
bracis-33576	193	10	and	and	CCONJ
bracis-33576	193	11	dbi	dbi	PROPN
bracis-33576	193	12	metrics	metric	NOUN
bracis-33576	193	13	to	to	PART
bracis-33576	193	14	assess	assess	VERB
bracis-33576	193	15	the	the	DET
bracis-33576	193	16	cohesion	cohesion	NOUN
bracis-33576	193	17	,	,	PUNCT
bracis-33576	193	18	compactness	compactness	NOUN
bracis-33576	193	19	,	,	PUNCT
bracis-33576	193	20	and	and	CCONJ
bracis-33576	193	21	separation	separation	NOUN
bracis-33576	193	22	of	of	ADP
bracis-33576	193	23	clusters	cluster	NOUN
bracis-33576	193	24	,	,	PUNCT
bracis-33576	193	25	ensuring	ensure	VERB
bracis-33576	193	26	a	a	DET
bracis-33576	193	27	robust	robust	ADJ
bracis-33576	193	28	analysis	analysis	NOUN
bracis-33576	193	29	of	of	ADP
bracis-33576	193	30	their	their	PRON
bracis-33576	193	31	structural	structural	ADJ
bracis-33576	193	32	integrity	integrity	NOUN
bracis-33576	193	33	.	.	PUNCT
bracis-33576	194	1	the	the	DET
bracis-33576	194	2	ss	ss	PROPN
bracis-33576	194	3	metric	metric	NOUN
bracis-33576	194	4	,	,	PUNCT
bracis-33576	194	5	which	which	PRON
bracis-33576	194	6	assesses	assess	VERB
bracis-33576	194	7	the	the	DET
bracis-33576	194	8	separation	separation	NOUN
bracis-33576	194	9	and	and	CCONJ
bracis-33576	194	10	cohesion	cohesion	NOUN
bracis-33576	194	11	of	of	ADP
bracis-33576	194	12	clusters	cluster	NOUN
bracis-33576	194	13	,	,	PUNCT
bracis-33576	194	14	is	be	AUX
bracis-33576	194	15	calculated	calculate	VERB
bracis-33576	194	16	for	for	ADP
bracis-33576	194	17	each	each	DET
bracis-33576	194	18	data	datum	NOUN
bracis-33576	194	19	point	point	NOUN
bracis-33576	194	20	\(i\	\(i\	PROPN
bracis-33576	194	21	)	)	PUNCT
bracis-33576	194	22	as	as	ADP
bracis-33576	194	23	:	:	PUNCT
bracis-33576	194	24	$	$	SYM
bracis-33576	194	25	$	$	SYM
bracis-33576	194	26	\begin{aligned	\begin{aligne	VERB
bracis-33576	194	27	}	}	PUNCT
bracis-33576	194	28	s(i	s(i	PROPN
bracis-33576	194	29	)	)	PUNCT
bracis-33576	194	30	=	=	SYM
bracis-33576	194	31	\frac{b(i	\frac{b(i	X
bracis-33576	194	32	)	)	PUNCT
bracis-33576	194	33	a(i)}{\max	a(i)}{\max	NOUN
bracis-33576	194	34	\{a(i	\{a(i	NOUN
bracis-33576	194	35	)	)	PUNCT
bracis-33576	194	36	,	,	PUNCT
bracis-33576	194	37	b(i)\	b(i)\	NOUN
bracis-33576	194	38	}	}	PUNCT
bracis-33576	194	39	}	}	PUNCT
bracis-33576	194	40	,	,	PUNCT
bracis-33576	194	41	\end{aligned}$$	\end{aligned}$$	VERB
bracis-33576	194	42	where	where	SCONJ
bracis-33576	194	43	\(a(i)\	\(a(i)\	NOUN
bracis-33576	194	44	)	)	PUNCT
bracis-33576	194	45	measures	measure	VERB
bracis-33576	194	46	the	the	DET
bracis-33576	194	47	average	average	ADJ
bracis-33576	194	48	intra	intra	ADJ
bracis-33576	194	49	-	-	ADJ
bracis-33576	194	50	cluster	cluster	NOUN
bracis-33576	194	51	distance	distance	NOUN
bracis-33576	194	52	,	,	PUNCT
bracis-33576	194	53	and	and	CCONJ
bracis-33576	194	54	\(b(i)\	\(b(i)\	NOUN
bracis-33576	194	55	)	)	PUNCT
bracis-33576	194	56	is	be	AUX
bracis-33576	194	57	the	the	DET
bracis-33576	194	58	minimum	minimum	ADJ
bracis-33576	194	59	inter	inter	ADJ
bracis-33576	194	60	-	-	ADJ
bracis-33576	194	61	cluster	cluster	ADJ
bracis-33576	194	62	distance	distance	NOUN
bracis-33576	194	63	for	for	ADP
bracis-33576	194	64	the	the	DET
bracis-33576	194	65	point	point	NOUN
bracis-33576	194	66	\(i\	\(i\	PROPN
bracis-33576	194	67	)	)	PUNCT
bracis-33576	194	68	.	.	PUNCT
bracis-33576	195	1	chi	chi	PROPN
bracis-33576	195	2	measures	measure	VERB
bracis-33576	195	3	the	the	DET
bracis-33576	195	4	ratio	ratio	NOUN
bracis-33576	195	5	of	of	ADP
bracis-33576	195	6	between	between	ADP
bracis-33576	195	7	-	-	PUNCT
bracis-33576	195	8	cluster	cluster	NOUN
bracis-33576	195	9	dispersion	dispersion	NOUN
bracis-33576	195	10	to	to	ADP
bracis-33576	195	11	within	within	ADP
bracis-33576	195	12	-	-	PUNCT
bracis-33576	195	13	cluster	cluster	NOUN
bracis-33576	195	14	dispersion	dispersion	NOUN
bracis-33576	195	15	,	,	PUNCT
bracis-33576	195	16	providing	provide	VERB
bracis-33576	195	17	insights	insight	NOUN
bracis-33576	195	18	into	into	ADP
bracis-33576	195	19	the	the	DET
bracis-33576	195	20	overall	overall	ADJ
bracis-33576	195	21	clustering	clustering	ADJ
bracis-33576	195	22	structure	structure	NOUN
bracis-33576	195	23	.	.	PUNCT
bracis-33576	196	1	the	the	DET
bracis-33576	196	2	index	index	NOUN
bracis-33576	196	3	is	be	AUX
bracis-33576	196	4	formulated	formulate	VERB
bracis-33576	196	5	as	as	ADP
bracis-33576	196	6	:	:	PUNCT
bracis-33576	196	7	$	$	SYM
bracis-33576	196	8	$	$	SYM
bracis-33576	196	9	\begin{aligned	\begin{aligne	VERB
bracis-33576	196	10	}	}	PUNCT
bracis-33576	196	11	chi	chi	NOUN
bracis-33576	196	12	=	=	PUNCT
bracis-33576	196	13	\frac{\text	\frac{\text	X
bracis-33576	196	14	{	{	PUNCT
bracis-33576	196	15	tr}(b_k	tr}(b_k	NOUN
bracis-33576	196	16	)	)	PUNCT
bracis-33576	196	17	/	/	SYM
bracis-33576	196	18	(	(	PUNCT
bracis-33576	196	19	k-1)}{\text	k-1)}{\text	PROPN
bracis-33576	196	20	{	{	PUNCT
bracis-33576	196	21	tr}(w_k	tr}(w_k	NOUN
bracis-33576	196	22	)	)	PUNCT
bracis-33576	196	23	/	/	SYM
bracis-33576	196	24	(	(	PUNCT
bracis-33576	196	25	n	n	CCONJ
bracis-33576	196	26	-	-	PUNCT
bracis-33576	196	27	k	k	NOUN
bracis-33576	196	28	)	)	PUNCT
bracis-33576	196	29	}	}	PUNCT
bracis-33576	196	30	,	,	PUNCT
bracis-33576	196	31	\end{aligned}$$	\end{aligned}$$	VERB
bracis-33576	196	32	where	where	SCONJ
bracis-33576	196	33	\(\text	\(\text	PROPN
bracis-33576	196	34	{	{	PUNCT
bracis-33576	196	35	tr}(b_k)\	tr}(b_k)\	NOUN
bracis-33576	196	36	)	)	PUNCT
bracis-33576	196	37	is	be	AUX
bracis-33576	196	38	the	the	DET
bracis-33576	196	39	trace	trace	NOUN
bracis-33576	196	40	of	of	ADP
bracis-33576	196	41	the	the	DET
bracis-33576	196	42	between	between	NOUN
bracis-33576	196	43	-	-	PUNCT
bracis-33576	196	44	group	group	NOUN
bracis-33576	196	45	dispersion	dispersion	NOUN
bracis-33576	196	46	matrix	matrix	NOUN
bracis-33576	196	47	,	,	PUNCT
bracis-33576	196	48	and	and	CCONJ
bracis-33576	196	49	\(\text	\(\text	NOUN
bracis-33576	196	50	{	{	PUNCT
bracis-33576	196	51	tr}(w_k)\	tr}(w_k)\	NOUN
bracis-33576	196	52	)	)	PUNCT
bracis-33576	196	53	the	the	DET
bracis-33576	196	54	trace	trace	NOUN
bracis-33576	196	55	of	of	ADP
bracis-33576	196	56	the	the	DET
bracis-33576	196	57	within	within	ADP
bracis-33576	196	58	-	-	PUNCT
bracis-33576	196	59	group	group	NOUN
bracis-33576	196	60	dispersion	dispersion	NOUN
bracis-33576	196	61	matrix	matrix	NOUN
bracis-33576	196	62	,	,	PUNCT
bracis-33576	196	63	thus	thus	ADV
bracis-33576	196	64	evaluating	evaluate	VERB
bracis-33576	196	65	both	both	CCONJ
bracis-33576	196	66	the	the	DET
bracis-33576	196	67	separation	separation	NOUN
bracis-33576	196	68	and	and	CCONJ
bracis-33576	196	69	compactness	compactness	NOUN
bracis-33576	196	70	of	of	ADP
bracis-33576	196	71	the	the	DET
bracis-33576	196	72	clusters	cluster	NOUN
bracis-33576	196	73	.	.	PUNCT
bracis-33576	197	1	the	the	DET
bracis-33576	197	2	dbi	dbi	PROPN
bracis-33576	197	3	evaluates	evaluate	VERB
bracis-33576	197	4	the	the	DET
bracis-33576	197	5	average	average	ADJ
bracis-33576	197	6	similarity	similarity	NOUN
bracis-33576	197	7	ratio	ratio	NOUN
bracis-33576	197	8	of	of	ADP
bracis-33576	197	9	each	each	DET
bracis-33576	197	10	cluster	cluster	NOUN
bracis-33576	197	11	with	with	ADP
bracis-33576	197	12	its	its	PRON
bracis-33576	197	13	most	most	ADV
bracis-33576	197	14	similar	similar	ADJ
bracis-33576	197	15	one	one	NOUN
bracis-33576	197	16	,	,	PUNCT
bracis-33576	197	17	offering	offer	VERB
bracis-33576	197	18	a	a	DET
bracis-33576	197	19	measure	measure	NOUN
bracis-33576	197	20	of	of	ADP
bracis-33576	197	21	cluster	cluster	NOUN
bracis-33576	197	22	separation	separation	NOUN
bracis-33576	197	23	.	.	PUNCT
bracis-33576	198	1	the	the	DET
bracis-33576	198	2	index	index	NOUN
bracis-33576	198	3	is	be	AUX
bracis-33576	198	4	calculated	calculate	VERB
bracis-33576	198	5	as	as	ADP
bracis-33576	198	6	:	:	PUNCT
bracis-33576	198	7	$	$	SYM
bracis-33576	198	8	$	$	SYM
bracis-33576	198	9	\begin{aligned	\begin{aligne	VERB
bracis-33576	198	10	}	}	PUNCT
bracis-33576	198	11	dbi	dbi	NOUN
bracis-33576	198	12	=	=	PUNCT
bracis-33576	198	13	\frac{1}{k	\frac{1}{k	NOUN
bracis-33576	198	14	}	}	PUNCT
bracis-33576	198	15	\sum	\sum	NOUN
bracis-33576	198	16	_	_	PUNCT
bracis-33576	198	17	{	{	PUNCT
bracis-33576	198	18	i=1}^{k	i=1}^{k	NOUN
bracis-33576	198	19	}	}	PUNCT
bracis-33576	198	20	\max	\max	PROPN
bracis-33576	198	21	_	_	PUNCT
bracis-33576	198	22	{	{	PUNCT
bracis-33576	198	23	j	j	PROPN
bracis-33576	198	24	\ne	\ne	PROPN
bracis-33576	198	25	i	i	PROPN
bracis-33576	198	26	}	}	PUNCT
bracis-33576	198	27	\left	\left	PROPN
bracis-33576	198	28	(	(	PUNCT
bracis-33576	198	29	\frac{\sigma	\frac{\sigma	NOUN
bracis-33576	198	30	_	_	PUNCT
bracis-33576	199	1	i	i	PRON
bracis-33576	199	2	+	+	X
bracis-33576	199	3	\sigma	\sigma	PROPN
bracis-33576	199	4	_	_	PUNCT
bracis-33576	199	5	j}{d_{ij	j}{d_{ij	NOUN
bracis-33576	199	6	}	}	PUNCT
bracis-33576	199	7	}	}	PUNCT
bracis-33576	199	8	\right	\right	PROPN
bracis-33576	199	9	)	)	PUNCT
bracis-33576	199	10	,	,	PUNCT
bracis-33576	199	11	\end{aligned}$$	\end{aligned}$$	VERB
bracis-33576	199	12	where	where	SCONJ
bracis-33576	199	13	\(\sigma	\(\sigma	NOUN
bracis-33576	199	14	_	_	PUNCT
bracis-33576	199	15	i\	i\	NOUN
bracis-33576	199	16	)	)	PUNCT
bracis-33576	199	17	and	and	CCONJ
bracis-33576	199	18	\(\sigma	\(\sigma	NOUN
bracis-33576	199	19	_	_	PUNCT
bracis-33576	199	20	j\	j\	VERB
bracis-33576	199	21	)	)	PUNCT
bracis-33576	199	22	represent	represent	VERB
bracis-33576	199	23	the	the	DET
bracis-33576	199	24	average	average	ADJ
bracis-33576	199	25	distance	distance	NOUN
bracis-33576	199	26	of	of	ADP
bracis-33576	199	27	all	all	DET
bracis-33576	199	28	elements	element	NOUN
bracis-33576	199	29	in	in	ADP
bracis-33576	199	30	clusters	cluster	NOUN
bracis-33576	199	31	\(i\	\(i\	PROPN
bracis-33576	199	32	)	)	PUNCT
bracis-33576	199	33	and	and	CCONJ
bracis-33576	199	34	\(j\	\(j\	NOUN
bracis-33576	199	35	)	)	PUNCT
bracis-33576	199	36	to	to	ADP
bracis-33576	199	37	their	their	PRON
bracis-33576	199	38	respective	respective	ADJ
bracis-33576	199	39	centroids	centroid	NOUN
bracis-33576	199	40	,	,	PUNCT
bracis-33576	199	41	and	and	CCONJ
bracis-33576	199	42	\(d_{ij}\	\(d_{ij}\	NOUN
bracis-33576	199	43	)	)	PUNCT
bracis-33576	199	44	is	be	AUX
bracis-33576	199	45	the	the	DET
bracis-33576	199	46	distance	distance	NOUN
bracis-33576	199	47	between	between	ADP
bracis-33576	199	48	centroids	centroid	NOUN
bracis-33576	199	49	of	of	ADP
bracis-33576	199	50	clusters	cluster	NOUN
bracis-33576	199	51	\(i\	\(i\	PROPN
bracis-33576	199	52	)	)	PUNCT
bracis-33576	199	53	and	and	CCONJ
bracis-33576	199	54	\(j\	\(j\	NOUN
bracis-33576	199	55	)	)	PUNCT
bracis-33576	199	56	.	.	PUNCT
bracis-33576	200	1	lower	low	ADJ
bracis-33576	200	2	values	value	NOUN
bracis-33576	200	3	of	of	ADP
bracis-33576	200	4	dbi	dbi	PROPN
bracis-33576	200	5	indicate	indicate	VERB
bracis-33576	200	6	better	well	ADJ
bracis-33576	200	7	cluster	cluster	NOUN
bracis-33576	200	8	separation	separation	NOUN
bracis-33576	200	9	.	.	PUNCT
bracis-33576	201	1	4.5	4.5	NUM
bracis-33576	201	2	our	our	PRON
bracis-33576	201	3	implemented	implement	VERB
bracis-33576	201	4	setup	setup	NOUN
bracis-33576	201	5	we	we	PRON
bracis-33576	201	6	compare	compare	VERB
bracis-33576	201	7	the	the	DET
bracis-33576	201	8	described	describe	VERB
bracis-33576	201	9	baselines	baseline	NOUN
bracis-33576	201	10	(	(	PUNCT
bracis-33576	201	11	cf	cf	NOUN
bracis-33576	201	12	.	.	PUNCT
bracis-33576	202	1	subsection	subsection	NOUN
bracis-33576	202	2	4.3	4.3	NUM
bracis-33576	202	3	)	)	PUNCT
bracis-33576	202	4	using	use	VERB
bracis-33576	202	5	a	a	DET
bracis-33576	202	6	pretrained	pretraine	VERB
bracis-33576	202	7	transformer	transformer	NOUN
bracis-33576	202	8	-	-	PUNCT
bracis-33576	202	9	decoder	decoder	NOUN
bracis-33576	202	10	llm	llm	PROPN
bracis-33576	202	11	model	model	NOUN
bracis-33576	202	12	,	,	PUNCT
bracis-33576	202	13	\(gpt-2\	\(gpt-2\	X
bracis-33576	202	14	)	)	PUNCT
bracis-33576	202	15	medium	medium	NOUN
bracis-33576	202	16	 	 	SPACE
bracis-33576	203	1	[	[	X
bracis-33576	203	2	16	16	NUM
bracis-33576	203	3	]	]	PUNCT
bracis-33576	203	4	,	,	PUNCT
bracis-33576	203	5	which	which	PRON
bracis-33576	203	6	has	have	VERB
bracis-33576	203	7	355	355	NUM
bracis-33576	203	8	million	million	NUM
bracis-33576	203	9	trainable	trainable	ADJ
bracis-33576	203	10	parameters	parameter	NOUN
bracis-33576	203	11	,	,	PUNCT
bracis-33576	203	12	24	24	NUM
bracis-33576	203	13	layers	layer	NOUN
bracis-33576	203	14	,	,	PUNCT
bracis-33576	203	15	16	16	NUM
bracis-33576	203	16	attention	attention	NOUN
bracis-33576	203	17	heads	head	NOUN
bracis-33576	203	18	,	,	PUNCT
bracis-33576	203	19	an	an	DET
bracis-33576	203	20	embedding	embed	VERB
bracis-33576	203	21	size	size	NOUN
bracis-33576	203	22	of	of	ADP
bracis-33576	203	23	1024	1024	NUM
bracis-33576	203	24	,	,	PUNCT
bracis-33576	203	25	and	and	CCONJ
bracis-33576	203	26	a	a	DET
bracis-33576	203	27	context	context	NOUN
bracis-33576	203	28	size	size	NOUN
bracis-33576	203	29	of	of	ADP
bracis-33576	203	30	1024	1024	NUM
bracis-33576	203	31	.	.	PUNCT
bracis-33576	204	1	to	to	PART
bracis-33576	204	2	facilitate	facilitate	VERB
bracis-33576	204	3	this	this	DET
bracis-33576	204	4	comparison	comparison	NOUN
bracis-33576	204	5	,	,	PUNCT
bracis-33576	204	6	we	we	PRON
bracis-33576	204	7	convert	convert	VERB
bracis-33576	204	8	the	the	DET
bracis-33576	204	9	tabular	tabular	NOUN
bracis-33576	204	10	dataset	dataset	VERB
bracis-33576	204	11	into	into	ADP
bracis-33576	204	12	text	text	NOUN
bracis-33576	204	13	for	for	ADP
bracis-33576	204	14	all	all	DET
bracis-33576	204	15	baselines	baseline	NOUN
bracis-33576	204	16	,	,	PUNCT
bracis-33576	204	17	applying	apply	VERB
bracis-33576	204	18	the	the	DET
bracis-33576	204	19	random	random	ADJ
bracis-33576	204	20	feature	feature	NOUN
bracis-33576	204	21	sequence	sequence	NOUN
bracis-33576	204	22	shuffling	shuffle	VERB
bracis-33576	204	23	function	function	NOUN
bracis-33576	204	24	and	and	CCONJ
bracis-33576	204	25	comparing	compare	VERB
bracis-33576	204	26	the	the	DET
bracis-33576	204	27	results	result	NOUN
bracis-33576	204	28	with	with	ADP
bracis-33576	204	29	erasmobase	erasmobase	NOUN
bracis-33576	204	30	and	and	CCONJ
bracis-33576	204	31	erasmonv	erasmonv	VERB
bracis-33576	204	32	.	.	PUNCT
bracis-33576	205	1	both	both	DET
bracis-33576	205	2	models	model	NOUN
bracis-33576	205	3	were	be	AUX
bracis-33576	205	4	trained	train	VERB
bracis-33576	205	5	with	with	ADP
bracis-33576	205	6	a	a	DET
bracis-33576	205	7	batch	batch	NOUN
bracis-33576	205	8	size	size	NOUN
bracis-33576	205	9	of	of	ADP
bracis-33576	205	10	8	8	NUM
bracis-33576	205	11	over	over	ADP
bracis-33576	205	12	60	60	NUM
bracis-33576	205	13	epochs	epoch	NOUN
bracis-33576	205	14	.	.	PUNCT
bracis-33576	206	1	we	we	PRON
bracis-33576	206	2	applied	apply	VERB
bracis-33576	206	3	a	a	DET
bracis-33576	206	4	dropout	dropout	NOUN
bracis-33576	206	5	rate	rate	NOUN
bracis-33576	206	6	of	of	ADP
bracis-33576	206	7	0.1	0.1	NUM
bracis-33576	206	8	and	and	CCONJ
bracis-33576	206	9	utilized	utilize	VERB
bracis-33576	206	10	500	500	NUM
bracis-33576	206	11	warmup	warmup	NOUN
bracis-33576	206	12	steps	step	NOUN
bracis-33576	206	13	.	.	PUNCT
bracis-33576	207	1	the	the	DET
bracis-33576	207	2	models	model	NOUN
bracis-33576	207	3	incorporated	incorporate	VERB
bracis-33576	207	4	a	a	DET
bracis-33576	207	5	weight	weight	NOUN
bracis-33576	207	6	decay	decay	NOUN
bracis-33576	207	7	of	of	ADP
bracis-33576	207	8	0.01	0.01	NUM
bracis-33576	207	9	and	and	CCONJ
bracis-33576	207	10	used	use	VERB
bracis-33576	207	11	the	the	DET
bracis-33576	207	12	adam	adam	PROPN
bracis-33576	207	13	optimizer	optimizer	NOUN
bracis-33576	207	14	with	with	ADP
bracis-33576	207	15	\(\epsilon	\(\epsilon	NOUN
bracis-33576	207	16	\	\	NOUN
bracis-33576	207	17	)	)	PUNCT
bracis-33576	207	18	set	set	VERB
bracis-33576	207	19	to	to	ADP
bracis-33576	207	20	1e–8	1e–8	NUM
bracis-33576	207	21	and	and	CCONJ
bracis-33576	207	22	\(\beta	\(\beta	PRON
bracis-33576	207	23	\	\	NOUN
bracis-33576	207	24	)	)	PUNCT
bracis-33576	207	25	values	value	NOUN
bracis-33576	207	26	of	of	ADP
bracis-33576	207	27	[	[	X
bracis-33576	207	28	0.7	0.7	NUM
bracis-33576	207	29	,	,	PUNCT
bracis-33576	207	30	0.9	0.9	NUM
bracis-33576	207	31	]	]	PUNCT
bracis-33576	207	32	.	.	PUNCT
bracis-33576	208	1	the	the	DET
bracis-33576	208	2	initial	initial	ADJ
bracis-33576	208	3	learning	learning	NOUN
bracis-33576	208	4	rate	rate	NOUN
bracis-33576	208	5	was	be	AUX
bracis-33576	208	6	set	set	VERB
bracis-33576	208	7	to	to	ADP
bracis-33576	208	8	5e–5	5e–5	NUM
bracis-33576	208	9	,	,	PUNCT
bracis-33576	208	10	with	with	ADP
bracis-33576	208	11	a	a	DET
bracis-33576	208	12	schedule	schedule	NOUN
bracis-33576	208	13	starting	start	VERB
bracis-33576	208	14	at	at	ADP
bracis-33576	208	15	1e–8	1e–8	NUM
bracis-33576	208	16	,	,	PUNCT
bracis-33576	208	17	ranging	range	VERB
bracis-33576	208	18	from	from	ADP
bracis-33576	208	19	a	a	DET
bracis-33576	208	20	minimum	minimum	NOUN
bracis-33576	208	21	of	of	ADP
bracis-33576	208	22	1e–5	1e–5	NUM
bracis-33576	208	23	to	to	ADP
bracis-33576	208	24	a	a	DET
bracis-33576	208	25	maximum	maximum	NOUN
bracis-33576	208	26	of	of	ADP
bracis-33576	208	27	4e-5	4e-5	PRON
bracis-33576	208	28	.	.	PUNCT
bracis-33576	209	1	the	the	DET
bracis-33576	209	2	model	model	NOUN
bracis-33576	209	3	was	be	AUX
bracis-33576	209	4	developed	develop	VERB
bracis-33576	209	5	using	use	VERB
bracis-33576	209	6	pytorchfootnote	pytorchfootnote	PROPN
bracis-33576	209	7	1	1	NUM
bracis-33576	209	8	and	and	CCONJ
bracis-33576	209	9	is	be	AUX
bracis-33576	209	10	made	make	VERB
bracis-33576	209	11	available	available	ADJ
bracis-33576	209	12	at	at	ADP
bracis-33576	209	13	a	a	DET
bracis-33576	209	14	github	github	NOUN
bracis-33576	209	15	repositoryfootnote	repositoryfootnote	NOUN
bracis-33576	209	16	2	2	NUM
bracis-33576	209	17	.	.	PUNCT
bracis-33576	210	1	it	it	PRON
bracis-33576	210	2	ran	run	VERB
bracis-33576	210	3	on	on	ADP
bracis-33576	210	4	a	a	DET
bracis-33576	210	5	system	system	NOUN
bracis-33576	210	6	equipped	equip	VERB
bracis-33576	210	7	with	with	ADP
bracis-33576	210	8	five	five	NUM
bracis-33576	210	9	nvidia	nvidia	PROPN
bracis-33576	210	10	rtx	rtx	PROPN
bracis-33576	210	11	a6000	a6000	PROPN
bracis-33576	210	12	,	,	PUNCT
bracis-33576	210	13	each	each	PRON
bracis-33576	210	14	having	have	VERB
bracis-33576	210	15	48	48	NUM
bracis-33576	210	16	gb	gb	NOUN
bracis-33576	210	17	of	of	ADP
bracis-33576	210	18	random	random	ADJ
bracis-33576	210	19	access	access	NOUN
bracis-33576	210	20	memory	memory	NOUN
bracis-33576	210	21	(	(	PUNCT
bracis-33576	210	22	ram	ram	NOUN
bracis-33576	210	23	)	)	PUNCT
bracis-33576	210	24	.	.	PUNCT
bracis-33576	211	1	5	5	NUM
bracis-33576	211	2	experimental	experimental	ADJ
bracis-33576	211	3	results	result	NOUN
bracis-33576	211	4	we	we	PRON
bracis-33576	211	5	present	present	VERB
bracis-33576	211	6	key	key	ADJ
bracis-33576	211	7	results	result	NOUN
bracis-33576	211	8	obtained	obtain	VERB
bracis-33576	211	9	organized	organize	VERB
bracis-33576	211	10	by	by	ADP
bracis-33576	211	11	datasets	dataset	NOUN
bracis-33576	211	12	.	.	PUNCT
bracis-33576	212	1	table	table	NOUN
bracis-33576	212	2	 	 	SPACE
bracis-33576	212	3	1	1	NUM
bracis-33576	212	4	presents	present	VERB
bracis-33576	212	5	the	the	DET
bracis-33576	212	6	ss	ss	NOUN
bracis-33576	212	7	,	,	PUNCT
bracis-33576	212	8	chi	chi	NOUN
bracis-33576	212	9	,	,	PUNCT
bracis-33576	212	10	and	and	CCONJ
bracis-33576	212	11	dbi	dbi	PROPN
bracis-33576	212	12	metrics	metric	NOUN
bracis-33576	212	13	results	result	VERB
bracis-33576	212	14	for	for	ADP
bracis-33576	212	15	the	the	DET
bracis-33576	212	16	test	test	NOUN
bracis-33576	212	17	dataset	dataset	VERB
bracis-33576	212	18	for	for	ADP
bracis-33576	212	19	all	all	DET
bracis-33576	212	20	evaluated	evaluated	ADJ
bracis-33576	212	21	approaches	approach	NOUN
bracis-33576	212	22	;	;	PUNCT
bracis-33576	212	23	values	value	NOUN
bracis-33576	212	24	in	in	ADP
bracis-33576	212	25	bold	bold	ADJ
bracis-33576	212	26	indicate	indicate	VERB
bracis-33576	212	27	the	the	DET
bracis-33576	212	28	best	good	ADJ
bracis-33576	212	29	outcomes	outcome	NOUN
bracis-33576	212	30	.	.	PUNCT
bracis-33576	213	1	the	the	DET
bracis-33576	213	2	best	good	ADJ
bracis-33576	213	3	algorithm	algorithm	NOUN
bracis-33576	213	4	was	be	AUX
bracis-33576	213	5	determined	determine	VERB
bracis-33576	213	6	by	by	ADP
bracis-33576	213	7	choosing	choose	VERB
bracis-33576	213	8	the	the	DET
bracis-33576	213	9	algorithm	algorithm	NOUN
bracis-33576	213	10	with	with	ADP
bracis-33576	213	11	the	the	DET
bracis-33576	213	12	highest	high	ADJ
bracis-33576	213	13	 	 	SPACE
bracis-33576	213	14	ss	ss	NOUN
bracis-33576	213	15	.	.	PUNCT
bracis-33576	214	1	banking	banking	NOUN
bracis-33576	214	2	.	.	PUNCT
bracis-33576	215	1	in	in	ADP
bracis-33576	215	2	the	the	DET
bracis-33576	215	3	banking	banking	NOUN
bracis-33576	215	4	dataset	dataset	NOUN
bracis-33576	215	5	,	,	PUNCT
bracis-33576	215	6	the	the	DET
bracis-33576	215	7	erasmobase	erasmobase	PROPN
bracis-33576	215	8	strategy	strategy	NOUN
bracis-33576	215	9	outperformed	outperform	VERB
bracis-33576	215	10	all	all	DET
bracis-33576	215	11	other	other	ADJ
bracis-33576	215	12	strategies	strategy	NOUN
bracis-33576	215	13	,	,	PUNCT
bracis-33576	215	14	achieving	achieve	VERB
bracis-33576	215	15	the	the	DET
bracis-33576	215	16	highest	high	ADJ
bracis-33576	215	17	 	 	SPACE
bracis-33576	215	18	ss	ss	NOUN
bracis-33576	215	19	of	of	ADP
bracis-33576	215	20	0.75	0.75	NUM
bracis-33576	215	21	,	,	PUNCT
bracis-33576	215	22	the	the	DET
bracis-33576	215	23	highest	high	ADJ
bracis-33576	215	24	chi	chi	NOUN
bracis-33576	215	25	of	of	ADP
bracis-33576	215	26	12	12	NUM
bracis-33576	215	27	,	,	PUNCT
bracis-33576	215	28	 	 	SPACE
bracis-33576	215	29	038.44	038.44	NUM
bracis-33576	215	30	,	,	PUNCT
bracis-33576	215	31	and	and	CCONJ
bracis-33576	215	32	the	the	DET
bracis-33576	215	33	lowest	low	ADJ
bracis-33576	215	34	dbi	dbi	NOUN
bracis-33576	215	35	of	of	ADP
bracis-33576	215	36	0.37	0.37	NUM
bracis-33576	215	37	.	.	PUNCT
bracis-33576	216	1	this	this	PRON
bracis-33576	216	2	indicates	indicate	VERB
bracis-33576	216	3	well	well	ADV
bracis-33576	216	4	-	-	PUNCT
bracis-33576	216	5	defined	define	VERB
bracis-33576	216	6	and	and	CCONJ
bracis-33576	216	7	compact	compact	ADJ
bracis-33576	216	8	clusters	cluster	NOUN
bracis-33576	216	9	.	.	PUNCT
bracis-33576	217	1	erasmonv	erasmonv	NOUN
bracis-33576	217	2	also	also	ADV
bracis-33576	217	3	performed	perform	VERB
bracis-33576	217	4	strongly	strongly	ADV
bracis-33576	217	5	with	with	ADP
bracis-33576	217	6	an	an	DET
bracis-33576	217	7	ss	ss	NOUN
bracis-33576	217	8	of	of	ADP
bracis-33576	217	9	0.71	0.71	NUM
bracis-33576	217	10	and	and	CCONJ
bracis-33576	217	11	chi	chi	NOUN
bracis-33576	217	12	of	of	ADP
bracis-33576	217	13	7	7	NUM
bracis-33576	217	14	,	,	PUNCT
bracis-33576	217	15	 	 	SPACE
bracis-33576	217	16	570.38	570.38	NUM
bracis-33576	217	17	,	,	PUNCT
bracis-33576	217	18	though	though	SCONJ
bracis-33576	217	19	it	it	PRON
bracis-33576	217	20	had	have	VERB
bracis-33576	217	21	a	a	DET
bracis-33576	217	22	slightly	slightly	ADV
bracis-33576	217	23	higher	high	ADJ
bracis-33576	217	24	dbi	dbi	NOUN
bracis-33576	217	25	of	of	ADP
bracis-33576	217	26	0.43	0.43	NUM
bracis-33576	217	27	compared	compare	VERB
bracis-33576	217	28	to	to	PART
bracis-33576	217	29	erasmobase	erasmobase	VERB
bracis-33576	217	30	.	.	PUNCT
bracis-33576	218	1	other	other	ADJ
bracis-33576	218	2	strategies	strategy	NOUN
bracis-33576	218	3	like	like	ADP
bracis-33576	218	4	mpnet	mpnet	NOUN
bracis-33576	218	5	-	-	PUNCT
bracis-33576	218	6	v2	v2	NOUN
bracis-33576	218	7	and	and	CCONJ
bracis-33576	218	8	falcon	falcon	NOUN
bracis-33576	218	9	showed	show	VERB
bracis-33576	218	10	moderate	moderate	ADJ
bracis-33576	218	11	performance	performance	NOUN
bracis-33576	218	12	with	with	ADP
bracis-33576	218	13	ss	ss	NOUN
bracis-33576	218	14	values	value	NOUN
bracis-33576	218	15	of	of	ADP
bracis-33576	218	16	0.27	0.27	NUM
bracis-33576	218	17	and	and	CCONJ
bracis-33576	218	18	0.23	0.23	NUM
bracis-33576	218	19	,	,	PUNCT
bracis-33576	218	20	respectively	respectively	ADV
bracis-33576	218	21	,	,	PUNCT
bracis-33576	218	22	while	while	SCONJ
bracis-33576	218	23	openai	openai	NOUN
bracis-33576	218	24	had	have	VERB
bracis-33576	218	25	the	the	DET
bracis-33576	218	26	lowest	low	ADJ
bracis-33576	218	27	ss	ss	NOUN
bracis-33576	218	28	at	at	ADP
bracis-33576	218	29	0.11	0.11	NUM
bracis-33576	218	30	and	and	CCONJ
bracis-33576	218	31	the	the	DET
bracis-33576	218	32	highest	high	ADJ
bracis-33576	218	33	dbi	dbi	NOUN
bracis-33576	218	34	at	at	ADP
bracis-33576	218	35	2.73	2.73	NUM
bracis-33576	218	36	,	,	PUNCT
bracis-33576	218	37	indicating	indicate	VERB
bracis-33576	218	38	poor	poor	ADJ
bracis-33576	218	39	clustering	clustering	NOUN
bracis-33576	218	40	.	.	PUNCT
bracis-33576	219	1	olist	olist	ADJ
bracis-33576	219	2	.	.	PUNCT
bracis-33576	220	1	for	for	ADP
bracis-33576	220	2	the	the	DET
bracis-33576	220	3	olist	olist	ADJ
bracis-33576	220	4	dataset	dataset	NOUN
bracis-33576	220	5	,	,	PUNCT
bracis-33576	220	6	erasmonv	erasmonv	VERB
bracis-33576	220	7	achieved	achieve	VERB
bracis-33576	220	8	the	the	DET
bracis-33576	220	9	highest	high	ADJ
bracis-33576	220	10	ss	ss	NOUN
bracis-33576	220	11	of	of	ADP
bracis-33576	220	12	0.77	0.77	NUM
bracis-33576	220	13	,	,	PUNCT
bracis-33576	220	14	indicating	indicate	VERB
bracis-33576	220	15	the	the	DET
bracis-33576	220	16	best	good	ADJ
bracis-33576	220	17	clustering	clustering	ADJ
bracis-33576	220	18	quality	quality	NOUN
bracis-33576	220	19	.	.	PUNCT
bracis-33576	221	1	it	it	PRON
bracis-33576	221	2	reached	reach	VERB
bracis-33576	221	3	the	the	DET
bracis-33576	221	4	highest	high	ADJ
bracis-33576	221	5	chi	chi	NOUN
bracis-33576	221	6	of	of	ADP
bracis-33576	221	7	62	62	NUM
bracis-33576	221	8	,	,	PUNCT
bracis-33576	221	9	 	 	SPACE
bracis-33576	221	10	036.87	036.87	NUM
bracis-33576	221	11	and	and	CCONJ
bracis-33576	221	12	a	a	DET
bracis-33576	221	13	low	low	ADJ
bracis-33576	221	14	dbi	dbi	NOUN
bracis-33576	221	15	of	of	ADP
bracis-33576	221	16	0.32	0.32	NUM
bracis-33576	221	17	.	.	PUNCT
bracis-33576	222	1	erasmobase	erasmobase	PROPN
bracis-33576	222	2	followed	follow	VERB
bracis-33576	222	3	closely	closely	ADV
bracis-33576	222	4	with	with	ADP
bracis-33576	222	5	an	an	DET
bracis-33576	222	6	ss	ss	NOUN
bracis-33576	222	7	of	of	ADP
bracis-33576	222	8	0.75	0.75	NUM
bracis-33576	222	9	and	and	CCONJ
bracis-33576	222	10	a	a	DET
bracis-33576	222	11	chi	chi	NOUN
bracis-33576	222	12	of	of	ADP
bracis-33576	222	13	54236.31	54236.31	NUM
bracis-33576	222	14	,	,	PUNCT
bracis-33576	222	15	along	along	ADP
bracis-33576	222	16	with	with	ADP
bracis-33576	222	17	the	the	DET
bracis-33576	222	18	lowest	low	ADJ
bracis-33576	222	19	dbi	dbi	NOUN
bracis-33576	222	20	of	of	ADP
bracis-33576	222	21	0.30	0.30	NUM
bracis-33576	222	22	.	.	PUNCT
bracis-33576	223	1	other	other	ADJ
bracis-33576	223	2	strategies	strategy	NOUN
bracis-33576	223	3	like	like	ADP
bracis-33576	223	4	llama-2	llama-2	NUM
bracis-33576	223	5	and	and	CCONJ
bracis-33576	223	6	falcon	falcon	PROPN
bracis-33576	223	7	performed	perform	VERB
bracis-33576	223	8	reasonably	reasonably	ADV
bracis-33576	223	9	well	well	ADV
bracis-33576	223	10	,	,	PUNCT
bracis-33576	223	11	with	with	ADP
bracis-33576	223	12	ss	ss	NOUN
bracis-33576	223	13	values	value	NOUN
bracis-33576	223	14	of	of	ADP
bracis-33576	223	15	0.71	0.71	NUM
bracis-33576	223	16	and	and	CCONJ
bracis-33576	223	17	0.66	0.66	NUM
bracis-33576	223	18	,	,	PUNCT
bracis-33576	223	19	respectively	respectively	ADV
bracis-33576	223	20	.	.	PUNCT
bracis-33576	224	1	however	however	ADV
bracis-33576	224	2	,	,	PUNCT
bracis-33576	224	3	openai	openai	NOUN
bracis-33576	224	4	and	and	CCONJ
bracis-33576	224	5	mpnet	mpnet	ADJ
bracis-33576	224	6	-	-	PUNCT
bracis-33576	224	7	v2	v2	NOUN
bracis-33576	224	8	showed	show	VERB
bracis-33576	224	9	lower	low	ADJ
bracis-33576	224	10	ss	ss	NOUN
bracis-33576	224	11	values	value	NOUN
bracis-33576	224	12	,	,	PUNCT
bracis-33576	224	13	with	with	ADP
bracis-33576	224	14	openai	openai	NOUN
bracis-33576	224	15	achieving	achieve	VERB
bracis-33576	224	16	an	an	DET
bracis-33576	224	17	ss	ss	NOUN
bracis-33576	224	18	of	of	ADP
bracis-33576	224	19	0.19	0.19	NUM
bracis-33576	224	20	and	and	CCONJ
bracis-33576	224	21	mpnet	mpnet	NOUN
bracis-33576	224	22	-	-	PUNCT
bracis-33576	224	23	v2	v2	NOUN
bracis-33576	224	24	an	an	DET
bracis-33576	224	25	ss	ss	NOUN
bracis-33576	224	26	of	of	ADP
bracis-33576	224	27	0.24	0.24	NUM
bracis-33576	224	28	.	.	PUNCT
bracis-33576	225	1	yelp	yelp	NOUN
bracis-33576	225	2	.	.	PUNCT
bracis-33576	226	1	in	in	ADP
bracis-33576	226	2	the	the	DET
bracis-33576	226	3	yelp	yelp	NOUN
bracis-33576	226	4	dataset	dataset	NOUN
bracis-33576	226	5	,	,	PUNCT
bracis-33576	226	6	erasmonv	erasmonv	VERB
bracis-33576	226	7	and	and	CCONJ
bracis-33576	226	8	erasmobase	erasmobase	PROPN
bracis-33576	226	9	both	both	PRON
bracis-33576	226	10	demonstrated	demonstrate	VERB
bracis-33576	226	11	superior	superior	ADJ
bracis-33576	226	12	performance	performance	NOUN
bracis-33576	226	13	,	,	PUNCT
bracis-33576	226	14	with	with	ADP
bracis-33576	226	15	ss	ss	NOUN
bracis-33576	226	16	values	value	NOUN
bracis-33576	226	17	of	of	ADP
bracis-33576	226	18	0.79	0.79	NUM
bracis-33576	226	19	and	and	CCONJ
bracis-33576	226	20	0.78	0.78	NUM
bracis-33576	226	21	,	,	PUNCT
bracis-33576	226	22	respectively	respectively	ADV
bracis-33576	226	23	.	.	PUNCT
bracis-33576	227	1	erasmonv	erasmonv	NOUN
bracis-33576	227	2	also	also	ADV
bracis-33576	227	3	achieved	achieve	VERB
bracis-33576	227	4	the	the	DET
bracis-33576	227	5	highest	high	ADJ
bracis-33576	227	6	chi	chi	NOUN
bracis-33576	227	7	of	of	ADP
bracis-33576	227	8	8	8	NUM
bracis-33576	227	9	,	,	PUNCT
bracis-33576	227	10	 	 	SPACE
bracis-33576	227	11	410.94	410.94	NUM
bracis-33576	227	12	and	and	CCONJ
bracis-33576	227	13	tied	tie	VERB
bracis-33576	227	14	with	with	ADP
bracis-33576	227	15	erasmobase	erasmobase	NOUN
bracis-33576	227	16	for	for	ADP
bracis-33576	227	17	the	the	DET
bracis-33576	227	18	lowest	low	ADJ
bracis-33576	227	19	dbi	dbi	NOUN
bracis-33576	227	20	of	of	ADP
bracis-33576	227	21	0.28	0.28	NUM
bracis-33576	227	22	.	.	PUNCT
bracis-33576	228	1	other	other	ADJ
bracis-33576	228	2	strategies	strategy	NOUN
bracis-33576	228	3	,	,	PUNCT
bracis-33576	228	4	such	such	ADJ
bracis-33576	228	5	as	as	ADP
bracis-33576	228	6	gpt2	gpt2	PROPN
bracis-33576	228	7	medium	medium	NOUN
bracis-33576	228	8	and	and	CCONJ
bracis-33576	228	9	llama-2	llama-2	NUM
bracis-33576	228	10	,	,	PUNCT
bracis-33576	228	11	showed	show	VERB
bracis-33576	228	12	moderate	moderate	ADJ
bracis-33576	228	13	clustering	clustering	NOUN
bracis-33576	228	14	performance	performance	NOUN
bracis-33576	228	15	with	with	ADP
bracis-33576	228	16	ss	ss	NOUN
bracis-33576	228	17	values	value	NOUN
bracis-33576	228	18	of	of	ADP
bracis-33576	228	19	0.39	0.39	NUM
bracis-33576	228	20	and	and	CCONJ
bracis-33576	228	21	0.29	0.29	NUM
bracis-33576	228	22	,	,	PUNCT
bracis-33576	228	23	respectively	respectively	ADV
bracis-33576	228	24	.	.	PUNCT
bracis-33576	229	1	openai	openai	PROPN
bracis-33576	229	2	,	,	PUNCT
bracis-33576	229	3	with	with	ADP
bracis-33576	229	4	an	an	DET
bracis-33576	229	5	ss	ss	NOUN
bracis-33576	229	6	of	of	ADP
bracis-33576	229	7	0.07	0.07	NUM
bracis-33576	229	8	,	,	PUNCT
bracis-33576	229	9	and	and	CCONJ
bracis-33576	229	10	mpnet	mpnet	NOUN
bracis-33576	229	11	-	-	PUNCT
bracis-33576	229	12	v2	v2	NOUN
bracis-33576	229	13	,	,	PUNCT
bracis-33576	229	14	with	with	ADP
bracis-33576	229	15	an	an	DET
bracis-33576	229	16	ss	ss	NOUN
bracis-33576	229	17	of	of	ADP
bracis-33576	229	18	0.23	0.23	NUM
bracis-33576	229	19	,	,	PUNCT
bracis-33576	229	20	indicated	indicate	VERB
bracis-33576	229	21	less	less	ADV
bracis-33576	229	22	effective	effective	ADJ
bracis-33576	229	23	clustering	clustering	NOUN
bracis-33576	229	24	.	.	PUNCT
bracis-33576	230	1	table	table	NOUN
bracis-33576	230	2	1	1	NUM
bracis-33576	230	3	.	.	PUNCT
bracis-33576	231	1	this	this	DET
bracis-33576	231	2	table	table	NOUN
bracis-33576	231	3	shows	show	VERB
bracis-33576	231	4	the	the	DET
bracis-33576	231	5	clustering	clustering	ADJ
bracis-33576	231	6	quality	quality	NOUN
bracis-33576	231	7	assessment	assessment	NOUN
bracis-33576	231	8	results	result	NOUN
bracis-33576	231	9	for	for	ADP
bracis-33576	231	10	the	the	DET
bracis-33576	231	11	top	top	ADV
bracis-33576	231	12	-	-	PUNCT
bracis-33576	231	13	performing	perform	VERB
bracis-33576	231	14	algorithms	algorithm	NOUN
bracis-33576	231	15	across	across	ADP
bracis-33576	231	16	each	each	DET
bracis-33576	231	17	dataset	dataset	NOUN
bracis-33576	231	18	and	and	CCONJ
bracis-33576	231	19	approach	approach	NOUN
bracis-33576	231	20	combination	combination	NOUN
bracis-33576	231	21	.	.	PUNCT
bracis-33576	232	1	the	the	DET
bracis-33576	232	2	best	good	ADJ
bracis-33576	232	3	algorithm	algorithm	NOUN
bracis-33576	232	4	was	be	AUX
bracis-33576	232	5	chosen	choose	VERB
bracis-33576	232	6	based	base	VERB
bracis-33576	232	7	on	on	ADP
bracis-33576	232	8	the	the	DET
bracis-33576	232	9	highest	high	ADJ
bracis-33576	232	10	ss	ss	NOUN
bracis-33576	232	11	.	.	PUNCT
bracis-33576	233	1	we	we	PRON
bracis-33576	233	2	provide	provide	VERB
bracis-33576	233	3	the	the	DET
bracis-33576	233	4	optimal	optimal	ADJ
bracis-33576	233	5	number	number	NOUN
bracis-33576	233	6	of	of	ADP
bracis-33576	233	7	clusters	cluster	NOUN
bracis-33576	233	8	(	(	PUNCT
bracis-33576	233	9	k	k	NOUN
bracis-33576	233	10	)	)	PUNCT
bracis-33576	233	11	and	and	CCONJ
bracis-33576	233	12	the	the	DET
bracis-33576	233	13	results	result	NOUN
bracis-33576	233	14	for	for	ADP
bracis-33576	233	15	ss	ss	NOUN
bracis-33576	233	16	,	,	PUNCT
bracis-33576	233	17	chi	chi	NOUN
bracis-33576	233	18	,	,	PUNCT
bracis-33576	233	19	and	and	CCONJ
bracis-33576	233	20	dbi	dbi	NOUN
bracis-33576	233	21	.	.	PUNCT
bracis-33576	234	1	bold	bold	ADJ
bracis-33576	234	2	values	value	NOUN
bracis-33576	234	3	indicate	indicate	VERB
bracis-33576	234	4	the	the	DET
bracis-33576	234	5	best	good	ADJ
bracis-33576	234	6	results	result	NOUN
bracis-33576	234	7	for	for	ADP
bracis-33576	234	8	each	each	DET
bracis-33576	234	9	metric.full	metric.full	NOUN
bracis-33576	234	10	size	size	NOUN
bracis-33576	234	11	table	table	NOUN
bracis-33576	234	12	petfinder.my	petfinder.my	PROPN
bracis-33576	234	13	.	.	PUNCT
bracis-33576	235	1	in	in	ADP
bracis-33576	235	2	the	the	DET
bracis-33576	235	3	petfinder.my	petfinder.my	PROPN
bracis-33576	235	4	dataset	dataset	NOUN
bracis-33576	235	5	,	,	PUNCT
bracis-33576	235	6	erasmonv	erasmonv	VERB
bracis-33576	235	7	slightly	slightly	ADV
bracis-33576	235	8	outperformed	outperform	VERB
bracis-33576	235	9	erasmobase	erasmobase	NOUN
bracis-33576	235	10	with	with	ADP
bracis-33576	235	11	an	an	DET
bracis-33576	235	12	ss	ss	NOUN
bracis-33576	235	13	of	of	ADP
bracis-33576	235	14	0.73	0.73	NUM
bracis-33576	235	15	compared	compare	VERB
bracis-33576	235	16	to	to	ADP
bracis-33576	235	17	0.72	0.72	NUM
bracis-33576	235	18	.	.	PUNCT
bracis-33576	236	1	erasmobase	erasmobase	PROPN
bracis-33576	236	2	had	have	VERB
bracis-33576	236	3	the	the	DET
bracis-33576	236	4	highest	high	ADJ
bracis-33576	236	5	chi	chi	NOUN
bracis-33576	236	6	of	of	ADP
bracis-33576	236	7	3	3	NUM
bracis-33576	236	8	,	,	PUNCT
bracis-33576	236	9	 	 	SPACE
bracis-33576	236	10	351.95	351.95	NUM
bracis-33576	236	11	and	and	CCONJ
bracis-33576	236	12	a	a	DET
bracis-33576	236	13	low	low	ADJ
bracis-33576	236	14	dbi	dbi	NOUN
bracis-33576	236	15	of	of	ADP
bracis-33576	236	16	0.40	0.40	NUM
bracis-33576	236	17	,	,	PUNCT
bracis-33576	236	18	while	while	SCONJ
bracis-33576	236	19	erasmonv	erasmonv	NOUN
bracis-33576	236	20	had	have	VERB
bracis-33576	236	21	a	a	DET
bracis-33576	236	22	chi	chi	NOUN
bracis-33576	236	23	of	of	ADP
bracis-33576	236	24	3	3	NUM
bracis-33576	236	25	,	,	PUNCT
bracis-33576	236	26	 	 	SPACE
bracis-33576	236	27	063.55	063.55	NUM
bracis-33576	236	28	and	and	CCONJ
bracis-33576	236	29	the	the	DET
bracis-33576	236	30	lowest	low	ADJ
bracis-33576	236	31	dbi	dbi	NOUN
bracis-33576	236	32	of	of	ADP
bracis-33576	236	33	0.34	0.34	NUM
bracis-33576	236	34	.	.	PUNCT
bracis-33576	237	1	other	other	ADJ
bracis-33576	237	2	strategies	strategy	NOUN
bracis-33576	237	3	,	,	PUNCT
bracis-33576	237	4	such	such	ADJ
bracis-33576	237	5	as	as	ADP
bracis-33576	237	6	gpt2	gpt2	PROPN
bracis-33576	237	7	medium	medium	NOUN
bracis-33576	237	8	and	and	CCONJ
bracis-33576	237	9	falcon	falcon	NOUN
bracis-33576	237	10	,	,	PUNCT
bracis-33576	237	11	showed	show	VERB
bracis-33576	237	12	moderate	moderate	ADJ
bracis-33576	237	13	results	result	NOUN
bracis-33576	237	14	with	with	ADP
bracis-33576	237	15	ss	ss	NOUN
bracis-33576	237	16	values	value	NOUN
bracis-33576	237	17	of	of	ADP
bracis-33576	237	18	0.55	0.55	NUM
bracis-33576	237	19	and	and	CCONJ
bracis-33576	237	20	0.20	0.20	NUM
bracis-33576	237	21	,	,	PUNCT
bracis-33576	237	22	respectively	respectively	ADV
bracis-33576	237	23	.	.	PUNCT
bracis-33576	238	1	mpnet	mpnet	NOUN
bracis-33576	238	2	-	-	PUNCT
bracis-33576	238	3	v2	v2	NOUN
bracis-33576	238	4	had	have	VERB
bracis-33576	238	5	the	the	DET
bracis-33576	238	6	lowest	low	ADJ
bracis-33576	238	7	ss	ss	NOUN
bracis-33576	238	8	of	of	ADP
bracis-33576	238	9	0.14	0.14	NUM
bracis-33576	238	10	,	,	PUNCT
bracis-33576	238	11	indicating	indicate	VERB
bracis-33576	238	12	poor	poor	ADJ
bracis-33576	238	13	clustering	clustering	ADJ
bracis-33576	238	14	performance	performance	NOUN
bracis-33576	238	15	.	.	PUNCT
bracis-33576	239	1	clothings	clothing	NOUN
bracis-33576	239	2	.	.	PUNCT
bracis-33576	240	1	for	for	SCONJ
bracis-33576	240	2	the	the	DET
bracis-33576	240	3	clothings	clothing	NOUN
bracis-33576	240	4	dataset	dataset	NOUN
bracis-33576	240	5	,	,	PUNCT
bracis-33576	240	6	erasmobase	erasmobase	PROPN
bracis-33576	240	7	achieved	achieve	VERB
bracis-33576	240	8	the	the	DET
bracis-33576	240	9	highest	high	ADJ
bracis-33576	240	10	ss	ss	NOUN
bracis-33576	240	11	of	of	ADP
bracis-33576	240	12	0.72	0.72	NUM
bracis-33576	240	13	and	and	CCONJ
bracis-33576	240	14	the	the	DET
bracis-33576	240	15	highest	high	ADJ
bracis-33576	240	16	chi	chi	NOUN
bracis-33576	240	17	of	of	ADP
bracis-33576	240	18	6	6	NUM
bracis-33576	240	19	,	,	PUNCT
bracis-33576	240	20	 	 	SPACE
bracis-33576	240	21	208.52	208.52	NUM
bracis-33576	240	22	,	,	PUNCT
bracis-33576	240	23	along	along	ADP
bracis-33576	240	24	with	with	ADP
bracis-33576	240	25	the	the	DET
bracis-33576	240	26	lowest	low	ADJ
bracis-33576	240	27	dbi	dbi	NOUN
bracis-33576	240	28	of	of	ADP
bracis-33576	240	29	0.39	0.39	NUM
bracis-33576	240	30	.	.	PUNCT
bracis-33576	241	1	erasmonv	erasmonv	PROPN
bracis-33576	241	2	also	also	ADV
bracis-33576	241	3	performed	perform	VERB
bracis-33576	241	4	well	well	ADV
bracis-33576	241	5	with	with	ADP
bracis-33576	241	6	an	an	DET
bracis-33576	241	7	ss	ss	NOUN
bracis-33576	241	8	of	of	ADP
bracis-33576	241	9	0.71	0.71	NUM
bracis-33576	241	10	and	and	CCONJ
bracis-33576	241	11	a	a	DET
bracis-33576	241	12	chi	chi	NOUN
bracis-33576	241	13	of	of	ADP
bracis-33576	241	14	5	5	NUM
bracis-33576	241	15	,	,	PUNCT
bracis-33576	241	16	 	 	SPACE
bracis-33576	241	17	916.57	916.57	NUM
bracis-33576	241	18	,	,	PUNCT
bracis-33576	241	19	matching	match	VERB
bracis-33576	241	20	the	the	DET
bracis-33576	241	21	lowest	low	ADJ
bracis-33576	241	22	dbi	dbi	NOUN
bracis-33576	241	23	of	of	ADP
bracis-33576	241	24	0.39	0.39	NUM
bracis-33576	241	25	.	.	PUNCT
bracis-33576	242	1	other	other	ADJ
bracis-33576	242	2	strategies	strategy	NOUN
bracis-33576	242	3	like	like	ADP
bracis-33576	242	4	gpt2	gpt2	PROPN
bracis-33576	242	5	medium	medium	NOUN
bracis-33576	242	6	and	and	CCONJ
bracis-33576	242	7	llama-2	llama-2	NUM
bracis-33576	242	8	showed	show	VERB
bracis-33576	242	9	moderate	moderate	ADJ
bracis-33576	242	10	performance	performance	NOUN
bracis-33576	242	11	,	,	PUNCT
bracis-33576	242	12	with	with	ADP
bracis-33576	242	13	ss	ss	NOUN
bracis-33576	242	14	values	value	NOUN
bracis-33576	242	15	of	of	ADP
bracis-33576	242	16	0.52	0.52	NUM
bracis-33576	242	17	and	and	CCONJ
bracis-33576	242	18	0.37	0.37	NUM
bracis-33576	242	19	,	,	PUNCT
bracis-33576	242	20	respectively	respectively	ADV
bracis-33576	242	21	.	.	PUNCT
bracis-33576	243	1	openai	openai	NOUN
bracis-33576	243	2	and	and	CCONJ
bracis-33576	243	3	mpnet	mpnet	NOUN
bracis-33576	243	4	-	-	PUNCT
bracis-33576	243	5	v2	v2	NOUN
bracis-33576	243	6	had	have	VERB
bracis-33576	243	7	the	the	DET
bracis-33576	243	8	lowest	low	ADJ
bracis-33576	243	9	ss	ss	NOUN
bracis-33576	243	10	values	value	NOUN
bracis-33576	243	11	of	of	ADP
bracis-33576	243	12	0.07	0.07	NUM
bracis-33576	243	13	and	and	CCONJ
bracis-33576	243	14	0.12	0.12	NUM
bracis-33576	243	15	,	,	PUNCT
bracis-33576	243	16	respectively	respectively	ADV
bracis-33576	243	17	,	,	PUNCT
bracis-33576	243	18	indicating	indicate	VERB
bracis-33576	243	19	poor	poor	ADJ
bracis-33576	243	20	clustering	clustering	ADJ
bracis-33576	243	21	quality	quality	NOUN
bracis-33576	243	22	.	.	PUNCT
bracis-33576	244	1	figure	figure	NOUN
bracis-33576	244	2	 	 	SPACE
bracis-33576	244	3	3	3	NUM
bracis-33576	244	4	shows	show	VERB
bracis-33576	244	5	a	a	DET
bracis-33576	244	6	2d	2d	NUM
bracis-33576	244	7	t	t	ADV
bracis-33576	244	8	-	-	PUNCT
bracis-33576	244	9	distributed	distribute	VERB
bracis-33576	244	10	stochastic	stochastic	ADJ
bracis-33576	244	11	neighbor	neighbor	NOUN
bracis-33576	244	12	embedding	embed	VERB
bracis-33576	244	13	(	(	PUNCT
bracis-33576	244	14	t	t	PROPN
bracis-33576	244	15	-	-	PUNCT
bracis-33576	244	16	sne	sne	NOUN
bracis-33576	244	17	)	)	PUNCT
bracis-33576	244	18	visualization	visualization	NOUN
bracis-33576	244	19	for	for	ADP
bracis-33576	244	20	the	the	DET
bracis-33576	244	21	yelp	yelp	NOUN
bracis-33576	244	22	dataset	dataset	NOUN
bracis-33576	244	23	across	across	ADP
bracis-33576	244	24	all	all	DET
bracis-33576	244	25	strategies	strategy	NOUN
bracis-33576	244	26	.	.	PUNCT
bracis-33576	245	1	the	the	DET
bracis-33576	245	2	yelp	yelp	NOUN
bracis-33576	245	3	dataset	dataset	NOUN
bracis-33576	245	4	was	be	AUX
bracis-33576	245	5	chosen	choose	VERB
bracis-33576	245	6	for	for	ADP
bracis-33576	245	7	its	its	PRON
bracis-33576	245	8	rich	rich	ADJ
bracis-33576	245	9	and	and	CCONJ
bracis-33576	245	10	diverse	diverse	ADJ
bracis-33576	245	11	user	user	NOUN
bracis-33576	245	12	reviews	review	NOUN
bracis-33576	245	13	,	,	PUNCT
bracis-33576	245	14	making	make	VERB
bracis-33576	245	15	it	it	PRON
bracis-33576	245	16	ideal	ideal	ADJ
bracis-33576	245	17	for	for	ADP
bracis-33576	245	18	clustering	cluster	VERB
bracis-33576	245	19	evaluation	evaluation	NOUN
bracis-33576	245	20	.	.	PUNCT
bracis-33576	246	1	our	our	PRON
bracis-33576	246	2	models	model	NOUN
bracis-33576	246	3	,	,	PUNCT
bracis-33576	246	4	erasmobase	erasmobase	NOUN
bracis-33576	246	5	and	and	CCONJ
bracis-33576	246	6	erasmonv	erasmonv	VERB
bracis-33576	246	7	,	,	PUNCT
bracis-33576	246	8	display	display	VERB
bracis-33576	246	9	distinct	distinct	ADJ
bracis-33576	246	10	and	and	CCONJ
bracis-33576	246	11	well	well	ADV
bracis-33576	246	12	-	-	PUNCT
bracis-33576	246	13	separated	separate	VERB
bracis-33576	246	14	clusters	cluster	NOUN
bracis-33576	246	15	,	,	PUNCT
bracis-33576	246	16	reflecting	reflect	VERB
bracis-33576	246	17	their	their	PRON
bracis-33576	246	18	lowest	low	ADJ
bracis-33576	246	19	dbi	dbi	NOUN
bracis-33576	246	20	scores	score	NOUN
bracis-33576	246	21	.	.	PUNCT
bracis-33576	247	1	this	this	DET
bracis-33576	247	2	highlights	highlight	VERB
bracis-33576	247	3	the	the	DET
bracis-33576	247	4	superior	superior	ADJ
bracis-33576	247	5	clustering	clustering	ADJ
bracis-33576	247	6	performance	performance	NOUN
bracis-33576	247	7	of	of	ADP
bracis-33576	247	8	our	our	PRON
bracis-33576	247	9	models	model	NOUN
bracis-33576	247	10	.	.	PUNCT
bracis-33576	248	1	fig	fig	NOUN
bracis-33576	248	2	.	.	PUNCT
bracis-33576	249	1	3	3	X
bracis-33576	249	2	.	.	X
bracis-33576	249	3	t	t	PROPN
bracis-33576	249	4	-	-	PUNCT
bracis-33576	249	5	sne	sne	NOUN
bracis-33576	249	6	visualization	visualization	NOUN
bracis-33576	249	7	of	of	ADP
bracis-33576	249	8	embedding	embed	VERB
bracis-33576	249	9	representations	representation	NOUN
bracis-33576	249	10	on	on	ADP
bracis-33576	249	11	the	the	DET
bracis-33576	249	12	yelp	yelp	NOUN
bracis-33576	249	13	dataset	dataset	NOUN
bracis-33576	249	14	for	for	ADP
bracis-33576	249	15	different	different	ADJ
bracis-33576	249	16	models	model	NOUN
bracis-33576	249	17	:	:	PUNCT
bracis-33576	249	18	(	(	PUNCT
bracis-33576	249	19	a	a	X
bracis-33576	249	20	)	)	PUNCT
bracis-33576	249	21	mpnet	mpnet	NOUN
bracis-33576	249	22	-	-	PUNCT
bracis-33576	249	23	v2	v2	NOUN
bracis-33576	249	24	,	,	PUNCT
bracis-33576	249	25	(	(	PUNCT
bracis-33576	249	26	b	b	NOUN
bracis-33576	249	27	)	)	PUNCT
bracis-33576	249	28	openai	openai	NOUN
bracis-33576	249	29	,	,	PUNCT
bracis-33576	249	30	(	(	PUNCT
bracis-33576	249	31	c	c	X
bracis-33576	249	32	)	)	PUNCT
bracis-33576	249	33	llama-2	llama-2	PROPN
bracis-33576	249	34	,	,	PUNCT
bracis-33576	249	35	(	(	PUNCT
bracis-33576	249	36	d	d	X
bracis-33576	249	37	)	)	PUNCT
bracis-33576	249	38	falcon	falcon	NOUN
bracis-33576	249	39	,	,	PUNCT
bracis-33576	249	40	(	(	PUNCT
bracis-33576	249	41	e	e	NOUN
bracis-33576	249	42	)	)	PUNCT
bracis-33576	249	43	gpt2	gpt2	PROPN
bracis-33576	249	44	medium	medium	NOUN
bracis-33576	249	45	,	,	PUNCT
bracis-33576	249	46	(	(	PUNCT
bracis-33576	249	47	f	f	X
bracis-33576	249	48	)	)	PUNCT
bracis-33576	249	49	pmv2	pmv2	NOUN
bracis-33576	249	50	+	+	CCONJ
bracis-33576	249	51	dice	dice	NOUN
bracis-33576	249	52	,	,	PUNCT
bracis-33576	249	53	(	(	PUNCT
bracis-33576	249	54	g	g	NOUN
bracis-33576	249	55	)	)	PUNCT
bracis-33576	249	56	erasmobase	erasmobase	NOUN
bracis-33576	249	57	,	,	PUNCT
bracis-33576	249	58	and	and	CCONJ
bracis-33576	249	59	(	(	PUNCT
bracis-33576	249	60	h	h	NOUN
bracis-33576	249	61	)	)	PUNCT
bracis-33576	249	62	erasmonv	erasmonv	VERB
bracis-33576	249	63	.	.	PUNCT
bracis-33576	250	1	full	full	ADJ
bracis-33576	250	2	size	size	NOUN
bracis-33576	250	3	image	image	NOUN
bracis-33576	250	4	6	6	NUM
bracis-33576	250	5	discussion	discussion	NOUN
bracis-33576	250	6	the	the	DET
bracis-33576	250	7	ss	ss	PROPN
bracis-33576	250	8	results	result	NOUN
bracis-33576	250	9	consistently	consistently	ADV
bracis-33576	250	10	highlight	highlight	VERB
bracis-33576	250	11	the	the	DET
bracis-33576	250	12	superior	superior	ADJ
bracis-33576	250	13	clustering	cluster	VERB
bracis-33576	250	14	effectiveness	effectiveness	NOUN
bracis-33576	250	15	of	of	ADP
bracis-33576	250	16	the	the	DET
bracis-33576	250	17	erasmobase	erasmobase	NOUN
bracis-33576	250	18	and	and	CCONJ
bracis-33576	250	19	erasmonv	erasmonv	ADJ
bracis-33576	250	20	strategies	strategy	NOUN
bracis-33576	250	21	across	across	ADP
bracis-33576	250	22	all	all	DET
bracis-33576	250	23	datasets	dataset	NOUN
bracis-33576	250	24	.	.	PUNCT
bracis-33576	251	1	erasmobase	erasmobase	PROPN
bracis-33576	251	2	achieved	achieve	VERB
bracis-33576	251	3	the	the	DET
bracis-33576	251	4	highest	high	ADJ
bracis-33576	251	5	ss	ss	NOUN
bracis-33576	251	6	in	in	ADP
bracis-33576	251	7	the	the	DET
bracis-33576	251	8	banking	banking	NOUN
bracis-33576	251	9	and	and	CCONJ
bracis-33576	251	10	clothing	clothing	NOUN
bracis-33576	251	11	datasets	dataset	NOUN
bracis-33576	251	12	,	,	PUNCT
bracis-33576	251	13	while	while	SCONJ
bracis-33576	251	14	erasmonv	erasmonv	NOUN
bracis-33576	251	15	led	lead	VERB
bracis-33576	251	16	in	in	ADP
bracis-33576	251	17	the	the	DET
bracis-33576	251	18	olist	olist	ADJ
bracis-33576	251	19	,	,	PUNCT
bracis-33576	251	20	yelp	yelp	NOUN
bracis-33576	251	21	,	,	PUNCT
bracis-33576	251	22	and	and	CCONJ
bracis-33576	251	23	petfinder.my	petfinder.my	PROPN
bracis-33576	251	24	datasets	dataset	NOUN
bracis-33576	251	25	.	.	PUNCT
bracis-33576	252	1	this	this	PRON
bracis-33576	252	2	indicates	indicate	VERB
bracis-33576	252	3	that	that	SCONJ
bracis-33576	252	4	both	both	DET
bracis-33576	252	5	strategies	strategy	NOUN
bracis-33576	252	6	help	help	VERB
bracis-33576	252	7	to	to	PART
bracis-33576	252	8	form	form	VERB
bracis-33576	252	9	well	well	ADV
bracis-33576	252	10	-	-	PUNCT
bracis-33576	252	11	defined	define	VERB
bracis-33576	252	12	clusters	cluster	NOUN
bracis-33576	252	13	,	,	PUNCT
bracis-33576	252	14	with	with	ADP
bracis-33576	252	15	erasmobase	erasmobase	NOUN
bracis-33576	252	16	having	have	VERB
bracis-33576	252	17	a	a	DET
bracis-33576	252	18	slight	slight	ADJ
bracis-33576	252	19	edge	edge	NOUN
bracis-33576	252	20	in	in	ADP
bracis-33576	252	21	some	some	DET
bracis-33576	252	22	datasets	dataset	NOUN
bracis-33576	252	23	.	.	PUNCT
bracis-33576	253	1	the	the	DET
bracis-33576	253	2	higher	high	ADJ
bracis-33576	253	3	ss	ss	NOUN
bracis-33576	253	4	values	value	NOUN
bracis-33576	253	5	for	for	ADP
bracis-33576	253	6	these	these	DET
bracis-33576	253	7	models	model	NOUN
bracis-33576	253	8	suggest	suggest	VERB
bracis-33576	253	9	they	they	PRON
bracis-33576	253	10	are	be	AUX
bracis-33576	253	11	better	well	ADJ
bracis-33576	253	12	at	at	ADP
bracis-33576	253	13	creating	create	VERB
bracis-33576	253	14	distinct	distinct	ADJ
bracis-33576	253	15	clusters	cluster	NOUN
bracis-33576	253	16	than	than	ADP
bracis-33576	253	17	other	other	ADJ
bracis-33576	253	18	strategies	strategy	NOUN
bracis-33576	253	19	like	like	ADP
bracis-33576	253	20	mpnet	mpnet	NOUN
bracis-33576	253	21	-	-	PUNCT
bracis-33576	253	22	v2	v2	NOUN
bracis-33576	253	23	,	,	PUNCT
bracis-33576	253	24	openai	openai	NOUN
bracis-33576	253	25	,	,	PUNCT
bracis-33576	253	26	and	and	CCONJ
bracis-33576	253	27	falcon	falcon	PROPN
bracis-33576	253	28	,	,	PUNCT
bracis-33576	253	29	which	which	PRON
bracis-33576	253	30	showed	show	VERB
bracis-33576	253	31	significantly	significantly	ADV
bracis-33576	253	32	lower	low	ADJ
bracis-33576	253	33	ss	ss	ADP
bracis-33576	253	34	values	value	NOUN
bracis-33576	253	35	,	,	PUNCT
bracis-33576	253	36	indicating	indicate	VERB
bracis-33576	253	37	poorer	poor	ADJ
bracis-33576	253	38	representation	representation	NOUN
bracis-33576	253	39	relevance	relevance	NOUN
bracis-33576	253	40	for	for	ADP
bracis-33576	253	41	clustering	cluster	VERB
bracis-33576	253	42	quality	quality	NOUN
bracis-33576	253	43	.	.	PUNCT
bracis-33576	254	1	the	the	DET
bracis-33576	254	2	dbi	dbi	PROPN
bracis-33576	254	3	results	result	VERB
bracis-33576	254	4	further	far	ADV
bracis-33576	254	5	support	support	VERB
bracis-33576	254	6	the	the	DET
bracis-33576	254	7	effectiveness	effectiveness	NOUN
bracis-33576	254	8	of	of	ADP
bracis-33576	254	9	erasmobase	erasmobase	NOUN
bracis-33576	254	10	and	and	CCONJ
bracis-33576	254	11	erasmonv	erasmonv	VERB
bracis-33576	254	12	.	.	PUNCT
bracis-33576	255	1	both	both	DET
bracis-33576	255	2	strategies	strategy	NOUN
bracis-33576	255	3	consistently	consistently	ADV
bracis-33576	255	4	achieved	achieve	VERB
bracis-33576	255	5	the	the	DET
bracis-33576	255	6	lowest	low	ADJ
bracis-33576	255	7	dbi	dbi	NOUN
bracis-33576	255	8	values	value	NOUN
bracis-33576	255	9	,	,	PUNCT
bracis-33576	255	10	particularly	particularly	ADV
bracis-33576	255	11	excelling	excel	VERB
bracis-33576	255	12	in	in	ADP
bracis-33576	255	13	the	the	DET
bracis-33576	255	14	banking	banking	NOUN
bracis-33576	255	15	,	,	PUNCT
bracis-33576	255	16	olist	olist	ADJ
bracis-33576	255	17	,	,	PUNCT
bracis-33576	255	18	and	and	CCONJ
bracis-33576	255	19	yelp	yelp	NOUN
bracis-33576	255	20	datasets	dataset	NOUN
bracis-33576	255	21	.	.	PUNCT
bracis-33576	256	1	a	a	DET
bracis-33576	256	2	lower	low	ADJ
bracis-33576	256	3	dbi	dbi	NOUN
bracis-33576	256	4	indicates	indicate	VERB
bracis-33576	256	5	more	more	ADV
bracis-33576	256	6	compact	compact	ADJ
bracis-33576	256	7	and	and	CCONJ
bracis-33576	256	8	well	well	ADV
bracis-33576	256	9	-	-	PUNCT
bracis-33576	256	10	separated	separate	VERB
bracis-33576	256	11	clusters	cluster	NOUN
bracis-33576	256	12	,	,	PUNCT
bracis-33576	256	13	affirming	affirm	VERB
bracis-33576	256	14	that	that	SCONJ
bracis-33576	256	15	our	our	PRON
bracis-33576	256	16	proposed	propose	VERB
bracis-33576	256	17	models	model	NOUN
bracis-33576	256	18	form	form	VERB
bracis-33576	256	19	tight	tight	ADJ
bracis-33576	256	20	and	and	CCONJ
bracis-33576	256	21	distinct	distinct	ADJ
bracis-33576	256	22	clusters	cluster	NOUN
bracis-33576	256	23	.	.	PUNCT
bracis-33576	257	1	erasmobase	erasmobase	PROPN
bracis-33576	257	2	generally	generally	ADV
bracis-33576	257	3	showed	show	VERB
bracis-33576	257	4	slightly	slightly	ADV
bracis-33576	257	5	better	well	ADJ
bracis-33576	257	6	dbi	dbi	PROPN
bracis-33576	257	7	scores	score	NOUN
bracis-33576	257	8	,	,	PUNCT
bracis-33576	257	9	suggesting	suggest	VERB
bracis-33576	257	10	it	it	PRON
bracis-33576	257	11	produces	produce	VERB
bracis-33576	257	12	more	more	ADJ
bracis-33576	257	13	compact	compact	ADJ
bracis-33576	257	14	clusters	cluster	NOUN
bracis-33576	257	15	than	than	ADP
bracis-33576	257	16	erasmonv	erasmonv	VERB
bracis-33576	257	17	.	.	PUNCT
bracis-33576	258	1	in	in	ADP
bracis-33576	258	2	contrast	contrast	NOUN
bracis-33576	258	3	,	,	PUNCT
bracis-33576	258	4	approaches	approach	VERB
bracis-33576	258	5	like	like	ADP
bracis-33576	258	6	openai	openai	NOUN
bracis-33576	258	7	and	and	CCONJ
bracis-33576	258	8	mpnet	mpnet	NOUN
bracis-33576	258	9	-	-	PUNCT
bracis-33576	258	10	v2	v2	NOUN
bracis-33576	258	11	had	have	VERB
bracis-33576	258	12	significantly	significantly	ADV
bracis-33576	258	13	higher	high	ADJ
bracis-33576	258	14	dbi	dbi	PROPN
bracis-33576	258	15	values	value	NOUN
bracis-33576	258	16	,	,	PUNCT
bracis-33576	258	17	reflecting	reflect	VERB
bracis-33576	258	18	less	less	ADV
bracis-33576	258	19	adequate	adequate	ADJ
bracis-33576	258	20	cluster	cluster	NOUN
bracis-33576	258	21	compactness	compactness	NOUN
bracis-33576	258	22	and	and	CCONJ
bracis-33576	258	23	separation	separation	NOUN
bracis-33576	258	24	.	.	PUNCT
bracis-33576	259	1	the	the	DET
bracis-33576	259	2	chi	chi	PROPN
bracis-33576	259	3	results	result	NOUN
bracis-33576	259	4	corroborate	corroborate	VERB
bracis-33576	259	5	the	the	DET
bracis-33576	259	6	trends	trend	NOUN
bracis-33576	259	7	observed	observe	VERB
bracis-33576	259	8	in	in	ADP
bracis-33576	259	9	ss	ss	NOUN
bracis-33576	259	10	and	and	CCONJ
bracis-33576	259	11	dbi	dbi	PROPN
bracis-33576	259	12	.	.	PUNCT
bracis-33576	260	1	erasmobase	erasmobase	PROPN
bracis-33576	260	2	and	and	CCONJ
bracis-33576	260	3	erasmonv	erasmonv	VERB
bracis-33576	260	4	achieved	achieve	VERB
bracis-33576	260	5	the	the	DET
bracis-33576	260	6	highest	high	ADJ
bracis-33576	260	7	chi	chi	NOUN
bracis-33576	260	8	scores	score	NOUN
bracis-33576	260	9	across	across	ADP
bracis-33576	260	10	most	most	ADJ
bracis-33576	260	11	datasets	dataset	NOUN
bracis-33576	260	12	,	,	PUNCT
bracis-33576	260	13	with	with	ADP
bracis-33576	260	14	erasmobase	erasmobase	NOUN
bracis-33576	260	15	leading	lead	VERB
bracis-33576	260	16	in	in	ADP
bracis-33576	260	17	the	the	DET
bracis-33576	260	18	banking	banking	NOUN
bracis-33576	260	19	and	and	CCONJ
bracis-33576	260	20	clothing	clothing	NOUN
bracis-33576	260	21	datasets	dataset	NOUN
bracis-33576	260	22	,	,	PUNCT
bracis-33576	260	23	and	and	CCONJ
bracis-33576	260	24	erasmonv	erasmonv	VERB
bracis-33576	260	25	in	in	ADP
bracis-33576	260	26	the	the	DET
bracis-33576	260	27	olist	olist	ADJ
bracis-33576	260	28	and	and	CCONJ
bracis-33576	260	29	yelp	yelp	NOUN
bracis-33576	260	30	datasets	dataset	NOUN
bracis-33576	260	31	.	.	PUNCT
bracis-33576	261	1	high	high	ADJ
bracis-33576	261	2	chi	chi	PROPN
bracis-33576	261	3	scores	score	NOUN
bracis-33576	261	4	indicate	indicate	VERB
bracis-33576	261	5	that	that	SCONJ
bracis-33576	261	6	our	our	PRON
bracis-33576	261	7	proposed	propose	VERB
bracis-33576	261	8	models	model	NOUN
bracis-33576	261	9	create	create	VERB
bracis-33576	261	10	clusters	cluster	NOUN
bracis-33576	261	11	that	that	PRON
bracis-33576	261	12	are	be	AUX
bracis-33576	261	13	not	not	PART
bracis-33576	261	14	only	only	ADV
bracis-33576	261	15	well	well	ADV
bracis-33576	261	16	-	-	PUNCT
bracis-33576	261	17	separated	separate	VERB
bracis-33576	261	18	but	but	CCONJ
bracis-33576	261	19	also	also	ADV
bracis-33576	261	20	highly	highly	ADV
bracis-33576	261	21	dense	dense	ADJ
bracis-33576	261	22	.	.	PUNCT
bracis-33576	262	1	these	these	DET
bracis-33576	262	2	superior	superior	ADJ
bracis-33576	262	3	results	result	NOUN
bracis-33576	262	4	across	across	ADP
bracis-33576	262	5	multiple	multiple	ADJ
bracis-33576	262	6	metrics	metric	NOUN
bracis-33576	262	7	underscores	underscore	VERB
bracis-33576	262	8	the	the	DET
bracis-33576	262	9	robustness	robustness	NOUN
bracis-33576	262	10	and	and	CCONJ
bracis-33576	262	11	effectiveness	effectiveness	NOUN
bracis-33576	262	12	of	of	ADP
bracis-33576	262	13	the	the	DET
bracis-33576	262	14	erasmo	erasmo	PROPN
bracis-33576	262	15	approach	approach	NOUN
bracis-33576	262	16	.	.	PUNCT
bracis-33576	263	1	other	other	ADJ
bracis-33576	263	2	models	model	NOUN
bracis-33576	263	3	like	like	ADP
bracis-33576	263	4	mpnet	mpnet	NOUN
bracis-33576	263	5	-	-	PUNCT
bracis-33576	263	6	v2	v2	NOUN
bracis-33576	263	7	,	,	PUNCT
bracis-33576	263	8	openai	openai	NOUN
bracis-33576	263	9	,	,	PUNCT
bracis-33576	263	10	and	and	CCONJ
bracis-33576	263	11	falcon	falcon	NOUN
bracis-33576	263	12	demonstrated	demonstrate	VERB
bracis-33576	263	13	lower	low	ADJ
bracis-33576	263	14	chi	chi	PROPN
bracis-33576	263	15	scores	score	NOUN
bracis-33576	263	16	,	,	PUNCT
bracis-33576	263	17	indicating	indicate	VERB
bracis-33576	263	18	less	less	ADV
bracis-33576	263	19	effective	effective	ADJ
bracis-33576	263	20	clustering	clustering	ADJ
bracis-33576	263	21	effectiveness	effectiveness	NOUN
bracis-33576	263	22	.	.	PUNCT
bracis-33576	264	1	erasmo	erasmo	PROPN
bracis-33576	264	2	has	have	VERB
bracis-33576	264	3	limitations	limitation	NOUN
bracis-33576	264	4	,	,	PUNCT
bracis-33576	264	5	including	include	VERB
bracis-33576	264	6	reliance	reliance	NOUN
bracis-33576	264	7	on	on	ADP
bracis-33576	264	8	high	high	ADJ
bracis-33576	264	9	-	-	PUNCT
bracis-33576	264	10	quality	quality	NOUN
bracis-33576	264	11	data	datum	NOUN
bracis-33576	264	12	,	,	PUNCT
bracis-33576	264	13	challenges	challenge	NOUN
bracis-33576	264	14	with	with	ADP
bracis-33576	264	15	ambiguous	ambiguous	ADJ
bracis-33576	264	16	datasets	dataset	NOUN
bracis-33576	264	17	,	,	PUNCT
bracis-33576	264	18	and	and	CCONJ
bracis-33576	264	19	significant	significant	ADJ
bracis-33576	264	20	computational	computational	ADJ
bracis-33576	264	21	demands	demand	NOUN
bracis-33576	264	22	for	for	ADP
bracis-33576	264	23	fine	fine	ADV
bracis-33576	264	24	-	-	PUNCT
bracis-33576	264	25	tuning	tuning	NOUN
bracis-33576	264	26	.	.	PUNCT
bracis-33576	265	1	random	random	ADJ
bracis-33576	265	2	sequence	sequence	NOUN
bracis-33576	265	3	shuffling	shuffle	VERB
bracis-33576	265	4	can	can	AUX
bracis-33576	265	5	affect	affect	VERB
bracis-33576	265	6	reproducibility	reproducibility	NOUN
bracis-33576	265	7	,	,	PUNCT
bracis-33576	265	8	and	and	CCONJ
bracis-33576	265	9	verbalizing	verbalize	VERB
bracis-33576	265	10	numerical	numerical	ADJ
bracis-33576	265	11	tokens	token	NOUN
bracis-33576	265	12	does	do	AUX
bracis-33576	265	13	not	not	PART
bracis-33576	265	14	consistently	consistently	ADV
bracis-33576	265	15	improve	improve	VERB
bracis-33576	265	16	outcomes	outcome	NOUN
bracis-33576	265	17	,	,	PUNCT
bracis-33576	265	18	warranting	warrant	VERB
bracis-33576	265	19	further	further	ADJ
bracis-33576	265	20	investigation	investigation	NOUN
bracis-33576	265	21	.	.	PUNCT
bracis-33576	266	1	furthermore	furthermore	ADV
bracis-33576	266	2	,	,	PUNCT
bracis-33576	266	3	applying	apply	VERB
bracis-33576	266	4	traditional	traditional	ADJ
bracis-33576	266	5	metrics	metric	NOUN
bracis-33576	266	6	such	such	ADJ
bracis-33576	266	7	as	as	ADP
bracis-33576	266	8	ss	ss	NOUN
bracis-33576	266	9	,	,	PUNCT
bracis-33576	266	10	chi	chi	NOUN
bracis-33576	266	11	,	,	PUNCT
bracis-33576	266	12	and	and	CCONJ
bracis-33576	266	13	dbi	dbi	NOUN
bracis-33576	266	14	to	to	PART
bracis-33576	266	15	analyze	analyze	VERB
bracis-33576	266	16	clusters	cluster	NOUN
bracis-33576	266	17	from	from	ADP
bracis-33576	266	18	various	various	ADJ
bracis-33576	266	19	representation	representation	NOUN
bracis-33576	266	20	spaces	space	NOUN
bracis-33576	266	21	appears	appear	VERB
bracis-33576	266	22	as	as	ADP
bracis-33576	266	23	an	an	DET
bracis-33576	266	24	open	open	ADJ
bracis-33576	266	25	research	research	NOUN
bracis-33576	266	26	challenge	challenge	NOUN
bracis-33576	266	27	yet	yet	ADV
bracis-33576	266	28	to	to	PART
bracis-33576	266	29	be	be	AUX
bracis-33576	266	30	resolved	resolve	VERB
bracis-33576	266	31	in	in	ADP
bracis-33576	266	32	the	the	DET
bracis-33576	266	33	literature	literature	NOUN
bracis-33576	266	34	.	.	PUNCT
bracis-33576	267	1	while	while	SCONJ
bracis-33576	267	2	these	these	DET
bracis-33576	267	3	metrics	metric	NOUN
bracis-33576	267	4	are	be	AUX
bracis-33576	267	5	designed	design	VERB
bracis-33576	267	6	for	for	ADP
bracis-33576	267	7	uniform	uniform	ADJ
bracis-33576	267	8	representation	representation	NOUN
bracis-33576	267	9	spaces	space	NOUN
bracis-33576	267	10	,	,	PUNCT
bracis-33576	267	11	they	they	PRON
bracis-33576	267	12	are	be	AUX
bracis-33576	267	13	frequently	frequently	ADV
bracis-33576	267	14	applied	apply	VERB
bracis-33576	267	15	to	to	ADP
bracis-33576	267	16	different	different	ADJ
bracis-33576	267	17	representation	representation	NOUN
bracis-33576	267	18	spaces	space	NOUN
bracis-33576	267	19	,	,	PUNCT
bracis-33576	267	20	as	as	SCONJ
bracis-33576	267	21	evidenced	evidence	VERB
bracis-33576	267	22	by	by	ADP
bracis-33576	267	23	previous	previous	ADJ
bracis-33576	267	24	studies	study	NOUN
bracis-33576	267	25	 	 	SPACE
bracis-33576	268	1	[	[	X
bracis-33576	268	2	6	6	NUM
bracis-33576	268	3	,	,	PUNCT
bracis-33576	268	4	14	14	NUM
bracis-33576	268	5	]	]	PUNCT
bracis-33576	268	6	.	.	PUNCT
bracis-33576	269	1	we	we	PRON
bracis-33576	269	2	recognize	recognize	VERB
bracis-33576	269	3	these	these	DET
bracis-33576	269	4	limitations	limitation	NOUN
bracis-33576	269	5	and	and	CCONJ
bracis-33576	269	6	stress	stress	VERB
bracis-33576	269	7	the	the	DET
bracis-33576	269	8	relevance	relevance	NOUN
bracis-33576	269	9	of	of	ADP
bracis-33576	269	10	investigating	investigate	VERB
bracis-33576	269	11	and	and	CCONJ
bracis-33576	269	12	developing	develop	VERB
bracis-33576	269	13	further	further	ADJ
bracis-33576	269	14	refined	refine	VERB
bracis-33576	269	15	metrics	metric	NOUN
bracis-33576	269	16	capable	capable	ADJ
bracis-33576	269	17	of	of	ADP
bracis-33576	269	18	accurately	accurately	ADV
bracis-33576	269	19	assessing	assess	VERB
bracis-33576	269	20	embeddings	embedding	NOUN
bracis-33576	269	21	from	from	ADP
bracis-33576	269	22	varied	varied	ADJ
bracis-33576	269	23	dimensional	dimensional	ADJ
bracis-33576	269	24	spaces	space	NOUN
bracis-33576	269	25	.	.	PUNCT
bracis-33576	270	1	despite	despite	SCONJ
bracis-33576	270	2	these	these	DET
bracis-33576	270	3	challenges	challenge	NOUN
bracis-33576	270	4	,	,	PUNCT
bracis-33576	270	5	erasmo	erasmo	PROPN
bracis-33576	270	6	significantly	significantly	ADV
bracis-33576	270	7	advances	advance	VERB
bracis-33576	270	8	clustering	cluster	VERB
bracis-33576	270	9	techniques	technique	NOUN
bracis-33576	270	10	by	by	ADP
bracis-33576	270	11	enhancing	enhance	VERB
bracis-33576	270	12	data	datum	NOUN
bracis-33576	270	13	representation	representation	NOUN
bracis-33576	270	14	.	.	PUNCT
bracis-33576	271	1	this	this	DET
bracis-33576	271	2	innovation	innovation	NOUN
bracis-33576	271	3	sets	set	VERB
bracis-33576	271	4	the	the	DET
bracis-33576	271	5	stage	stage	NOUN
bracis-33576	271	6	for	for	ADP
bracis-33576	271	7	the	the	DET
bracis-33576	271	8	potential	potential	ADJ
bracis-33576	271	9	standardization	standardization	NOUN
bracis-33576	271	10	of	of	ADP
bracis-33576	271	11	metrics	metric	NOUN
bracis-33576	271	12	and	and	CCONJ
bracis-33576	271	13	improves	improve	VERB
bracis-33576	271	14	clustering	cluster	VERB
bracis-33576	271	15	quality	quality	NOUN
bracis-33576	271	16	.	.	PUNCT
bracis-33576	272	1	erasmo	erasmo	PROPN
bracis-33576	272	2	stands	stand	VERB
bracis-33576	272	3	as	as	ADP
bracis-33576	272	4	a	a	DET
bracis-33576	272	5	crucial	crucial	ADJ
bracis-33576	272	6	development	development	NOUN
bracis-33576	272	7	in	in	ADP
bracis-33576	272	8	pursuing	pursue	VERB
bracis-33576	272	9	more	more	ADV
bracis-33576	272	10	robust	robust	ADJ
bracis-33576	272	11	clustering	clustering	ADJ
bracis-33576	272	12	strategies	strategy	NOUN
bracis-33576	272	13	.	.	PUNCT
bracis-33576	273	1	7	7	NUM
bracis-33576	273	2	conclusion	conclusion	NOUN
bracis-33576	273	3	exploring	explore	VERB
bracis-33576	273	4	structured	structure	VERB
bracis-33576	273	5	and	and	CCONJ
bracis-33576	273	6	textual	textual	ADJ
bracis-33576	273	7	data	datum	NOUN
bracis-33576	273	8	simultaneously	simultaneously	ADV
bracis-33576	273	9	for	for	ADP
bracis-33576	273	10	clustering	cluster	VERB
bracis-33576	273	11	analysis	analysis	NOUN
bracis-33576	273	12	is	be	AUX
bracis-33576	273	13	a	a	DET
bracis-33576	273	14	challenging	challenging	ADJ
bracis-33576	273	15	problem	problem	NOUN
bracis-33576	273	16	.	.	PUNCT
bracis-33576	274	1	this	this	DET
bracis-33576	274	2	study	study	NOUN
bracis-33576	274	3	presented	present	VERB
bracis-33576	274	4	and	and	CCONJ
bracis-33576	274	5	evaluated	evaluate	VERB
bracis-33576	274	6	the	the	DET
bracis-33576	274	7	erasmobase	erasmobase	NOUN
bracis-33576	274	8	and	and	CCONJ
bracis-33576	274	9	erasmonv	erasmonv	VERB
bracis-33576	274	10	clustering	clustering	ADJ
bracis-33576	274	11	approaches	approach	NOUN
bracis-33576	274	12	,	,	PUNCT
bracis-33576	274	13	demonstrating	demonstrate	VERB
bracis-33576	274	14	their	their	PRON
bracis-33576	274	15	superior	superior	ADJ
bracis-33576	274	16	empirical	empirical	ADJ
bracis-33576	274	17	effectiveness	effectiveness	NOUN
bracis-33576	274	18	across	across	ADP
bracis-33576	274	19	multiple	multiple	ADJ
bracis-33576	274	20	datasets	dataset	NOUN
bracis-33576	274	21	.	.	PUNCT
bracis-33576	275	1	the	the	DET
bracis-33576	275	2	results	result	NOUN
bracis-33576	275	3	,	,	PUNCT
bracis-33576	275	4	based	base	VERB
bracis-33576	275	5	on	on	ADP
bracis-33576	275	6	ss	ss	PROPN
bracis-33576	275	7	,	,	PUNCT
bracis-33576	275	8	dbi	dbi	PROPN
bracis-33576	275	9	,	,	PUNCT
bracis-33576	275	10	and	and	CCONJ
bracis-33576	275	11	chi	chi	NOUN
bracis-33576	275	12	,	,	PUNCT
bracis-33576	275	13	consistently	consistently	ADV
bracis-33576	275	14	revealed	reveal	VERB
bracis-33576	275	15	that	that	SCONJ
bracis-33576	275	16	our	our	PRON
bracis-33576	275	17	proposed	propose	VERB
bracis-33576	275	18	models	model	NOUN
bracis-33576	275	19	help	help	VERB
bracis-33576	275	20	to	to	PART
bracis-33576	275	21	create	create	VERB
bracis-33576	275	22	well	well	ADV
bracis-33576	275	23	-	-	PUNCT
bracis-33576	275	24	defined	define	VERB
bracis-33576	275	25	,	,	PUNCT
bracis-33576	275	26	compact	compact	ADJ
bracis-33576	275	27	,	,	PUNCT
bracis-33576	275	28	and	and	CCONJ
bracis-33576	275	29	dense	dense	ADJ
bracis-33576	275	30	clusters	cluster	NOUN
bracis-33576	275	31	,	,	PUNCT
bracis-33576	275	32	outperforming	outperform	VERB
bracis-33576	275	33	all	all	DET
bracis-33576	275	34	strategies	strategy	NOUN
bracis-33576	275	35	.	.	PUNCT
bracis-33576	276	1	despite	despite	SCONJ
bracis-33576	276	2	their	their	PRON
bracis-33576	276	3	high	high	ADJ
bracis-33576	276	4	computational	computational	ADJ
bracis-33576	276	5	demands	demand	NOUN
bracis-33576	276	6	and	and	CCONJ
bracis-33576	276	7	dependency	dependency	NOUN
bracis-33576	276	8	on	on	ADP
bracis-33576	276	9	data	datum	NOUN
bracis-33576	276	10	quality	quality	NOUN
bracis-33576	276	11	,	,	PUNCT
bracis-33576	276	12	we	we	PRON
bracis-33576	276	13	found	find	VERB
bracis-33576	276	14	the	the	DET
bracis-33576	276	15	erasmo	erasmo	PROPN
bracis-33576	276	16	models	model	NOUN
bracis-33576	276	17	suited	suit	VERB
bracis-33576	276	18	for	for	ADP
bracis-33576	276	19	practical	practical	ADJ
bracis-33576	276	20	clustering	clustering	NOUN
bracis-33576	276	21	tools	tool	NOUN
bracis-33576	276	22	.	.	PUNCT
bracis-33576	277	1	future	future	ADJ
bracis-33576	277	2	work	work	NOUN
bracis-33576	277	3	might	might	AUX
bracis-33576	277	4	focus	focus	VERB
bracis-33576	277	5	on	on	ADP
bracis-33576	277	6	optimizing	optimize	VERB
bracis-33576	277	7	computational	computational	ADJ
bracis-33576	277	8	efficiency	efficiency	NOUN
bracis-33576	277	9	and	and	CCONJ
bracis-33576	277	10	enhancing	enhance	VERB
bracis-33576	277	11	robustness	robustness	NOUN
bracis-33576	277	12	for	for	ADP
bracis-33576	277	13	diverse	diverse	ADJ
bracis-33576	277	14	and	and	CCONJ
bracis-33576	277	15	noisy	noisy	ADJ
bracis-33576	277	16	datasets	dataset	NOUN
bracis-33576	277	17	.	.	PUNCT
bracis-33576	278	1	notes	note	VERB
bracis-33576	278	2	1.pytorch.org	1.pytorch.org	NUM
bracis-33576	278	3	.	.	PUNCT
bracis-33576	278	4	2.erasmo	2.erasmo	NUM
bracis-33576	278	5	github	github	NOUN
bracis-33576	278	6	.	.	PUNCT
bracis-33576	279	1	references	reference	NOUN
bracis-33576	279	2	almazrouei	almazrouei	PROPN
bracis-33576	279	3	,	,	PUNCT
bracis-33576	279	4	e.	e.	PROPN
bracis-33576	279	5	,	,	PUNCT
bracis-33576	279	6	et	et	PROPN
bracis-33576	279	7	 	 	SPACE
bracis-33576	279	8	al	al	PROPN
bracis-33576	279	9	.	.	PROPN
bracis-33576	279	10	:	:	PUNCT
bracis-33576	280	1	falcon-40b	falcon-40b	ADV
bracis-33576	280	2	:	:	PUNCT
bracis-33576	280	3	an	an	DET
bracis-33576	280	4	open	open	ADJ
bracis-33576	280	5	large	large	ADJ
bracis-33576	280	6	language	language	NOUN
bracis-33576	280	7	model	model	NOUN
bracis-33576	280	8	with	with	ADP
bracis-33576	280	9	state	state	NOUN
bracis-33576	280	10	-	-	PUNCT
bracis-33576	280	11	of	of	ADP
bracis-33576	280	12	-	-	PUNCT
bracis-33576	280	13	the	the	DET
bracis-33576	280	14	-	-	PUNCT
bracis-33576	280	15	art	art	NOUN
bracis-33576	280	16	performance	performance	NOUN
bracis-33576	280	17	(	(	PUNCT
bracis-33576	280	18	2023	2023	NUM
bracis-33576	280	19	)	)	PUNCT
bracis-33576	280	20	google	google	NOUN
bracis-33576	280	21	scholar	scholar	NOUN
bracis-33576	280	22	  	  	SPACE
bracis-33576	280	23	borisov	borisov	NOUN
bracis-33576	280	24	,	,	PUNCT
bracis-33576	280	25	v.	v.	ADV
bracis-33576	280	26	,	,	PUNCT
bracis-33576	280	27	seßler	seßler	PROPN
bracis-33576	280	28	,	,	PUNCT
bracis-33576	280	29	k.	k.	PROPN
bracis-33576	280	30	,	,	PUNCT
bracis-33576	280	31	leemann	leemann	PROPN
bracis-33576	280	32	,	,	PUNCT
bracis-33576	280	33	t.	t.	PROPN
bracis-33576	280	34	,	,	PUNCT
bracis-33576	280	35	pawelczyk	pawelczyk	NOUN
bracis-33576	280	36	,	,	PUNCT
bracis-33576	280	37	m.	m.	NOUN
bracis-33576	280	38	,	,	PUNCT
bracis-33576	280	39	kasneci	kasneci	PROPN
bracis-33576	280	40	,	,	PUNCT
bracis-33576	280	41	g.	g.	NOUN
bracis-33576	280	42	:	:	PUNCT
bracis-33576	280	43	language	language	NOUN
bracis-33576	280	44	models	model	NOUN
bracis-33576	280	45	are	be	AUX
bracis-33576	280	46	realistic	realistic	ADJ
bracis-33576	280	47	tabular	tabular	NOUN
bracis-33576	280	48	data	datum	NOUN
bracis-33576	280	49	generators	generator	NOUN
bracis-33576	280	50	.	.	PUNCT
bracis-33576	281	1	arxiv	arxiv	PROPN
bracis-33576	281	2	preprint	preprint	VERB
bracis-33576	281	3	arxiv:2210.06280	arxiv:2210.06280	PROPN
bracis-33576	281	4	(	(	PUNCT
bracis-33576	281	5	2022	2022	NUM
bracis-33576	281	6	)	)	PUNCT
bracis-33576	281	7	brooks	brook	NOUN
bracis-33576	281	8	,	,	PUNCT
bracis-33576	281	9	n.	n.	NOUN
bracis-33576	281	10	:	:	PUNCT
bracis-33576	281	11	women	woman	NOUN
bracis-33576	281	12	’s	’s	PART
bracis-33576	281	13	e	e	PROPN
bracis-33576	281	14	-	-	ADJ
bracis-33576	281	15	commerce	commerce	ADJ
bracis-33576	281	16	clothing	clothing	NOUN
bracis-33576	281	17	reviews	review	NOUN
bracis-33576	281	18	(	(	PUNCT
bracis-33576	281	19	2017	2017	NUM
bracis-33576	281	20	)	)	PUNCT
bracis-33576	281	21	.	.	PUNCT
bracis-33576	282	1	https://www.kaggle.com/datasets/nicapotato/womens-ecommerce-clothing-reviews	https://www.kaggle.com/datasets/nicapotato/womens-ecommerce-clothing-reviews	PROPN
bracis-33576	282	2	dataset	dataset	PROPN
bracis-33576	282	3	,	,	PUNCT
bracis-33576	282	4	y.	y.	PROPN
bracis-33576	282	5	:	:	PUNCT
bracis-33576	282	6	yelp	yelp	NOUN
bracis-33576	282	7	dataset	dataset	NOUN
bracis-33576	282	8	(	(	PUNCT
bracis-33576	282	9	2014	2014	NUM
bracis-33576	282	10	)	)	PUNCT
bracis-33576	282	11	.	.	PUNCT
bracis-33576	283	1	http://www.yelp.com/dataset_challenge	http://www.yelp.com/dataset_challenge	PROPN
bracis-33576	283	2	kaggle	kaggle	PROPN
bracis-33576	283	3	,	,	PUNCT
bracis-33576	283	4	petfinder.my	petfinder.my	PROPN
bracis-33576	283	5	:	:	PUNCT
bracis-33576	283	6	petfinder.my	petfinder.my	PROPN
bracis-33576	283	7	adoption	adoption	NOUN
bracis-33576	283	8	prediction	prediction	NOUN
bracis-33576	283	9	(	(	PUNCT
bracis-33576	283	10	2019	2019	NUM
bracis-33576	283	11	)	)	PUNCT
bracis-33576	283	12	.	.	PUNCT
bracis-33576	284	1	https://www.kaggle.com/c/petfinder-adoption-prediction/data/	https://www.kaggle.com/c/petfinder-adoption-prediction/data/	PROPN
bracis-33576	284	2	keraghel	keraghel	PROPN
bracis-33576	284	3	,	,	PUNCT
bracis-33576	284	4	i.	i.	PROPN
bracis-33576	284	5	,	,	PUNCT
bracis-33576	284	6	et	et	PROPN
bracis-33576	284	7	 	 	SPACE
bracis-33576	284	8	al	al	PROPN
bracis-33576	284	9	.	.	PROPN
bracis-33576	284	10	:	:	PUNCT
bracis-33576	285	1	beyond	beyond	ADP
bracis-33576	285	2	words	word	NOUN
bracis-33576	285	3	:	:	PUNCT
bracis-33576	285	4	a	a	DET
bracis-33576	285	5	comparative	comparative	ADJ
bracis-33576	285	6	analysis	analysis	NOUN
bracis-33576	285	7	of	of	ADP
bracis-33576	285	8	llm	llm	PROPN
bracis-33576	285	9	embeddings	embedding	NOUN
bracis-33576	285	10	for	for	ADP
bracis-33576	285	11	effective	effective	ADJ
bracis-33576	285	12	clustering	clustering	NOUN
bracis-33576	285	13	.	.	PUNCT
bracis-33576	286	1	in	in	ADP
bracis-33576	286	2	:	:	PUNCT
bracis-33576	286	3	int	int	NOUN
bracis-33576	286	4	.	.	PUNCT
bracis-33576	287	1	symposium	symposium	NOUN
bracis-33576	287	2	on	on	ADP
bracis-33576	287	3	intelligent	intelligent	ADJ
bracis-33576	287	4	data	datum	NOUN
bracis-33576	287	5	analysis	analysis	NOUN
bracis-33576	287	6	,	,	PUNCT
bracis-33576	287	7	pp	pp	X
bracis-33576	287	8	.	.	PUNCT
bracis-33576	288	1	205–216	205–216	NUM
bracis-33576	288	2	.	.	X
bracis-33576	288	3	springer	springer	NOUN
bracis-33576	288	4	,	,	PUNCT
bracis-33576	288	5	heidelberg	heidelberg	PROPN
bracis-33576	288	6	(	(	PUNCT
bracis-33576	288	7	2024	2024	NUM
bracis-33576	288	8	)	)	PUNCT
bracis-33576	288	9	google	google	PROPN
bracis-33576	288	10	scholar	scholar	NOUN
bracis-33576	288	11	  	  	SPACE
bracis-33576	288	12	lewis	lewis	PROPN
bracis-33576	288	13	,	,	PUNCT
bracis-33576	288	14	p.	p.	PROPN
bracis-33576	288	15	,	,	PUNCT
bracis-33576	288	16	perez	perez	PROPN
bracis-33576	288	17	,	,	PUNCT
bracis-33576	288	18	a.	a.	NOUN
bracis-33576	288	19	,	,	PUNCT
bracis-33576	288	20	petroni	petroni	NOUN
bracis-33576	288	21	,	,	PUNCT
bracis-33576	288	22	w.t	w.t	PROPN
bracis-33576	288	23	.	.	PROPN
bracis-33576	288	24	,	,	PUNCT
bracis-33576	288	25	rocktäschel	rocktäschel	PROPN
bracis-33576	288	26	,	,	PUNCT
bracis-33576	288	27	t.	t.	PROPN
bracis-33576	288	28	,	,	PUNCT
bracis-33576	288	29	et	et	PROPN
bracis-33576	288	30	 	 	SPACE
bracis-33576	288	31	al	al	PROPN
bracis-33576	288	32	.	.	PROPN
bracis-33576	288	33	:	:	PUNCT
bracis-33576	289	1	retrieval	retrieval	NOUN
bracis-33576	289	2	-	-	PUNCT
bracis-33576	289	3	augmented	augment	VERB
bracis-33576	289	4	generation	generation	NOUN
bracis-33576	289	5	for	for	ADP
bracis-33576	289	6	knowledge	knowledge	NOUN
bracis-33576	289	7	-	-	PUNCT
bracis-33576	289	8	intensive	intensive	ADJ
bracis-33576	289	9	nlp	nlp	NOUN
bracis-33576	289	10	tasks	task	NOUN
bracis-33576	289	11	.	.	PUNCT
bracis-33576	290	1	in	in	ADP
bracis-33576	290	2	:	:	PUNCT
bracis-33576	290	3	nips	nip	NOUN
bracis-33576	290	4	(	(	PUNCT
bracis-33576	290	5	2020	2020	NUM
bracis-33576	290	6	)	)	PUNCT
bracis-33576	290	7	google	google	PROPN
bracis-33576	290	8	scholar	scholar	NOUN
bracis-33576	290	9	  	  	SPACE
bracis-33576	290	10	liu	liu	PROPN
bracis-33576	290	11	,	,	PUNCT
bracis-33576	290	12	y.	y.	PROPN
bracis-33576	290	13	,	,	PUNCT
bracis-33576	290	14	he	he	PRON
bracis-33576	290	15	,	,	PUNCT
bracis-33576	290	16	x.	x.	PROPN
bracis-33576	290	17	,	,	PUNCT
bracis-33576	290	18	zhong	zhong	PROPN
bracis-33576	290	19	,	,	PUNCT
bracis-33576	290	20	t.	t.	PROPN
bracis-33576	290	21	,	,	PUNCT
bracis-33576	290	22	et	et	PROPN
bracis-33576	290	23	 	 	SPACE
bracis-33576	290	24	al	al	PROPN
bracis-33576	290	25	.	.	PROPN
bracis-33576	290	26	:	:	PUNCT
bracis-33576	290	27	understanding	understand	VERB
bracis-33576	290	28	llms	llm	NOUN
bracis-33576	290	29	:	:	PUNCT
bracis-33576	290	30	a	a	DET
bracis-33576	290	31	comprehensive	comprehensive	ADJ
bracis-33576	290	32	overview	overview	NOUN
bracis-33576	290	33	from	from	ADP
bracis-33576	290	34	training	training	NOUN
bracis-33576	290	35	to	to	ADP
bracis-33576	290	36	inference	inference	NOUN
bracis-33576	290	37	.	.	PUNCT
bracis-33576	291	1	arxiv	arxiv	PROPN
bracis-33576	291	2	preprint	preprint	PROPN
bracis-33576	291	3	arxiv:2401.02038	arxiv:2401.02038	PROPN
bracis-33576	291	4	(	(	PUNCT
bracis-33576	291	5	2024	2024	NUM
bracis-33576	291	6	)	)	PUNCT
bracis-33576	291	7	min	min	PROPN
bracis-33576	291	8	,	,	PUNCT
bracis-33576	291	9	b.	b.	PROPN
bracis-33576	291	10	,	,	PUNCT
bracis-33576	291	11	ross	ross	PROPN
bracis-33576	291	12	,	,	PUNCT
bracis-33576	291	13	d.	d.	PROPN
bracis-33576	291	14	,	,	PUNCT
bracis-33576	291	15	others	other	NOUN
bracis-33576	291	16	.	.	PUNCT
bracis-33576	291	17	:	:	PUNCT
bracis-33576	292	1	recent	recent	ADJ
bracis-33576	292	2	advances	advance	NOUN
bracis-33576	292	3	in	in	ADP
bracis-33576	292	4	natural	natural	ADJ
bracis-33576	292	5	language	language	NOUN
bracis-33576	292	6	processing	processing	NOUN
bracis-33576	292	7	via	via	ADP
bracis-33576	292	8	large	large	ADJ
bracis-33576	292	9	pre	pre	ADJ
bracis-33576	292	10	-	-	ADJ
bracis-33576	292	11	trained	train	VERB
bracis-33576	292	12	language	language	NOUN
bracis-33576	292	13	models	model	NOUN
bracis-33576	292	14	:	:	PUNCT
bracis-33576	292	15	a	a	DET
bracis-33576	292	16	survey	survey	NOUN
bracis-33576	292	17	.	.	PUNCT
bracis-33576	293	1	acm	acm	PROPN
bracis-33576	293	2	comput	comput	NOUN
bracis-33576	293	3	.	.	PUNCT
bracis-33576	294	1	surv	surv	NOUN
bracis-33576	294	2	.	.	PUNCT
bracis-33576	295	1	1–40	1–40	PROPN
bracis-33576	295	2	(	(	PUNCT
bracis-33576	295	3	2023	2023	NUM
bracis-33576	295	4	)	)	PUNCT
bracis-33576	295	5	google	google	NOUN
bracis-33576	295	6	scholar	scholar	NOUN
bracis-33576	295	7	  	  	SPACE
bracis-33576	295	8	moro	moro	PROPN
bracis-33576	295	9	,	,	PUNCT
bracis-33576	295	10	s.	s.	PROPN
bracis-33576	295	11	 	 	SPACE
bracis-33576	295	12	rita	rita	PROPN
bracis-33576	295	13	,	,	PUNCT
bracis-33576	295	14	p.	p.	PROPN
bracis-33576	295	15	,	,	PUNCT
bracis-33576	295	16	cortez	cortez	PROPN
bracis-33576	295	17	,	,	PUNCT
bracis-33576	295	18	p.	p.	NOUN
bracis-33576	295	19	:	:	PUNCT
bracis-33576	295	20	bank	bank	NOUN
bracis-33576	295	21	marketing	marketing	NOUN
bracis-33576	295	22	.	.	PUNCT
bracis-33576	296	1	uci	uci	PROPN
bracis-33576	296	2	machine	machine	NOUN
bracis-33576	296	3	learning	learn	VERB
bracis-33576	296	4	repository	repository	NOUN
bracis-33576	296	5	(	(	PUNCT
bracis-33576	296	6	2012	2012	NUM
bracis-33576	296	7	)	)	PUNCT
bracis-33576	296	8	.	.	PUNCT
bracis-33576	297	1	https://doi.org/10.24432/c5k306	https://doi.org/10.24432/c5k306	PROPN
bracis-33576	297	2	olist	olist	ADJ
bracis-33576	297	3	:	:	PUNCT
bracis-33576	297	4	brazilian	brazilian	ADJ
bracis-33576	297	5	e	e	PROPN
bracis-33576	297	6	-	-	ADJ
bracis-33576	297	7	commerce	commerce	ADJ
bracis-33576	297	8	public	public	NOUN
bracis-33576	297	9	dataset	dataset	VERB
bracis-33576	297	10	by	by	ADP
bracis-33576	297	11	olist	olist	ADJ
bracis-33576	297	12	(	(	PUNCT
bracis-33576	297	13	2023	2023	NUM
bracis-33576	297	14	)	)	PUNCT
bracis-33576	297	15	.	.	PUNCT
bracis-33576	298	1	https://doi.org/10.34740/kaggle/dsv/195341	https://doi.org/10.34740/kaggle/dsv/195341	PROPN
bracis-33576	298	2	.	.	PUNCT
bracis-33576	299	1	acessado	acessado	VERB
bracis-33576	299	2	em	em	PRON
bracis-33576	299	3	:	:	PUNCT
bracis-33576	299	4	03	03	NUM
bracis-33576	299	5	de	de	X
bracis-33576	299	6	outubro	outubro	X
bracis-33576	299	7	de	de	PROPN
bracis-33576	299	8	2023	2023	NUM
bracis-33576	299	9	ourabah	ourabah	NOUN
bracis-33576	299	10	,	,	PUNCT
bracis-33576	299	11	o.	o.	PROPN
bracis-33576	299	12	,	,	PUNCT
bracis-33576	299	13	et	et	PROPN
bracis-33576	299	14	al	al	PROPN
bracis-33576	299	15	.	.	PUNCT
bracis-33576	299	16	:	:	PUNCT
bracis-33576	300	1	large	large	ADJ
bracis-33576	300	2	scale	scale	NOUN
bracis-33576	300	3	data	datum	NOUN
bracis-33576	300	4	using	use	VERB
bracis-33576	300	5	k	k	NOUN
bracis-33576	300	6	-	-	PUNCT
bracis-33576	300	7	means	mean	NOUN
bracis-33576	300	8	.	.	PUNCT
bracis-33576	301	1	mesopotamian	mesopotamian	PROPN
bracis-33576	301	2	j.	j.	PROPN
bracis-33576	301	3	big	big	PROPN
bracis-33576	301	4	data	datum	NOUN
bracis-33576	301	5	2023	2023	NUM
bracis-33576	301	6	,	,	PUNCT
bracis-33576	301	7	36–45	36–45	NUM
bracis-33576	301	8	(	(	PUNCT
bracis-33576	301	9	2023	2023	NUM
bracis-33576	301	10	)	)	PUNCT
bracis-33576	301	11	math	math	NOUN
bracis-33576	301	12	  	  	SPACE
bracis-33576	301	13	google	google	PROPN
bracis-33576	301	14	scholar	scholar	NOUN
bracis-33576	301	15	  	  	SPACE
bracis-33576	301	16	pedregosa	pedregosa	PROPN
bracis-33576	301	17	,	,	PUNCT
bracis-33576	301	18	f.	f.	PROPN
bracis-33576	301	19	,	,	PUNCT
bracis-33576	301	20	varoquaux	varoquaux	PROPN
bracis-33576	301	21	,	,	PUNCT
bracis-33576	301	22	m.	m.	NOUN
bracis-33576	301	23	,	,	PUNCT
bracis-33576	301	24	prettenhofer	prettenhofer	NOUN
bracis-33576	301	25	,	,	PUNCT
bracis-33576	301	26	v.	v.	PROPN
bracis-33576	301	27	,	,	PUNCT
bracis-33576	301	28	et	et	NOUN
bracis-33576	301	29	 	 	SPACE
bracis-33576	301	30	al	al	PROPN
bracis-33576	301	31	.	.	PROPN
bracis-33576	301	32	:	:	PUNCT
bracis-33576	302	1	scikit	scikit	NOUN
bracis-33576	302	2	-	-	PUNCT
bracis-33576	302	3	learn	learn	VERB
bracis-33576	302	4	:	:	PUNCT
bracis-33576	302	5	machine	machine	NOUN
bracis-33576	302	6	learning	learning	NOUN
bracis-33576	302	7	in	in	ADP
bracis-33576	302	8	python	python	PROPN
bracis-33576	302	9	.	.	PUNCT
bracis-33576	303	1	j.	j.	PROPN
bracis-33576	303	2	mach	mach	PROPN
bracis-33576	303	3	.	.	PUNCT
bracis-33576	304	1	learn	learn	VERB
bracis-33576	304	2	.	.	PUNCT
bracis-33576	305	1	res	re	NOUN
bracis-33576	305	2	.	.	PROPN
bracis-33576	306	1	12	12	NUM
bracis-33576	306	2	,	,	PUNCT
bracis-33576	306	3	2825–2830	2825–2830	NUM
bracis-33576	306	4	(	(	PUNCT
bracis-33576	306	5	2011	2011	NUM
bracis-33576	306	6	)	)	PUNCT
bracis-33576	306	7	google	google	PROPN
bracis-33576	306	8	scholar	scholar	NOUN
bracis-33576	306	9	  	  	SPACE
bracis-33576	306	10	petukhova	petukhova	PROPN
bracis-33576	306	11	,	,	PUNCT
bracis-33576	306	12	a.	a.	NOUN
bracis-33576	306	13	,	,	PUNCT
bracis-33576	306	14	matos	matos	PROPN
bracis-33576	306	15	-	-	NOUN
bracis-33576	306	16	carvalho	carvalho	NOUN
bracis-33576	306	17	,	,	PUNCT
bracis-33576	306	18	j.p	j.p	PROPN
bracis-33576	306	19	.	.	PROPN
bracis-33576	306	20	,	,	PUNCT
bracis-33576	306	21	fachada	fachada	PROPN
bracis-33576	306	22	,	,	PUNCT
bracis-33576	306	23	n.	n.	NOUN
bracis-33576	306	24	:	:	PUNCT
bracis-33576	306	25	text	text	NOUN
bracis-33576	306	26	clustering	cluster	VERB
bracis-33576	306	27	with	with	ADP
bracis-33576	306	28	llm	llm	PROPN
bracis-33576	306	29	embeddings	embedding	NOUN
bracis-33576	306	30	.	.	PUNCT
bracis-33576	307	1	arxiv	arxiv	PROPN
bracis-33576	307	2	preprint	preprint	VERB
bracis-33576	307	3	arxiv:2403.15112	arxiv:2403.15112	PROPN
bracis-33576	307	4	(	(	PUNCT
bracis-33576	307	5	2024	2024	NUM
bracis-33576	307	6	)	)	PUNCT
bracis-33576	307	7	pitafi	pitafi	NOUN
bracis-33576	307	8	,	,	PUNCT
bracis-33576	307	9	s.	s.	PROPN
bracis-33576	307	10	,	,	PUNCT
bracis-33576	307	11	anwar	anwar	PROPN
bracis-33576	307	12	,	,	PUNCT
bracis-33576	307	13	t.	t.	PROPN
bracis-33576	307	14	,	,	PUNCT
bracis-33576	307	15	sharif	sharif	PROPN
bracis-33576	307	16	,	,	PUNCT
bracis-33576	307	17	z.	z.	PROPN
bracis-33576	307	18	:	:	PUNCT
bracis-33576	307	19	a	a	DET
bracis-33576	307	20	taxonomy	taxonomy	NOUN
bracis-33576	307	21	of	of	ADP
bracis-33576	307	22	machine	machine	NOUN
bracis-33576	307	23	learning	learn	VERB
bracis-33576	307	24	clustering	cluster	VERB
bracis-33576	307	25	algorithms	algorithm	NOUN
bracis-33576	307	26	,	,	PUNCT
bracis-33576	307	27	challenges	challenge	NOUN
bracis-33576	307	28	,	,	PUNCT
bracis-33576	307	29	and	and	CCONJ
bracis-33576	307	30	future	future	ADJ
bracis-33576	307	31	realms	realm	NOUN
bracis-33576	307	32	.	.	PUNCT
bracis-33576	308	1	appl	appl	PROPN
bracis-33576	308	2	.	.	PUNCT
bracis-33576	309	1	sci	sci	PROPN
bracis-33576	309	2	.	.	PROPN
bracis-33576	309	3	13(6	13(6	PROPN
bracis-33576	309	4	)	)	PUNCT
bracis-33576	309	5	,	,	PUNCT
bracis-33576	309	6	3529	3529	NUM
bracis-33576	309	7	(	(	PUNCT
bracis-33576	309	8	2023	2023	NUM
bracis-33576	309	9	)	)	PUNCT
bracis-33576	309	10	article	article	NOUN
bracis-33576	309	11	  	  	SPACE
bracis-33576	309	12	math	math	NOUN
bracis-33576	309	13	  	  	SPACE
bracis-33576	309	14	google	google	PROPN
bracis-33576	309	15	scholar	scholar	NOUN
bracis-33576	309	16	  	  	SPACE
bracis-33576	309	17	radford	radford	PROPN
bracis-33576	309	18	,	,	PUNCT
bracis-33576	309	19	a.	a.	PROPN
bracis-33576	309	20	,	,	PUNCT
bracis-33576	309	21	wu	wu	PROPN
bracis-33576	309	22	,	,	PUNCT
bracis-33576	309	23	j.	j.	PROPN
bracis-33576	309	24	,	,	PUNCT
bracis-33576	309	25	child	child	NOUN
bracis-33576	309	26	,	,	PUNCT
bracis-33576	309	27	r.	r.	PROPN
bracis-33576	309	28	,	,	PUNCT
bracis-33576	309	29	luan	luan	PROPN
bracis-33576	309	30	,	,	PUNCT
bracis-33576	309	31	d.	d.	PROPN
bracis-33576	309	32	,	,	PUNCT
bracis-33576	309	33	amodei	amodei	PROPN
bracis-33576	309	34	,	,	PUNCT
bracis-33576	309	35	d.	d.	PROPN
bracis-33576	309	36	,	,	PUNCT
bracis-33576	309	37	sutskever	sutskever	PROPN
bracis-33576	309	38	,	,	PUNCT
bracis-33576	309	39	i.	i.	PROPN
bracis-33576	309	40	,	,	PUNCT
bracis-33576	309	41	et	et	PROPN
bracis-33576	309	42	al	al	PROPN
bracis-33576	309	43	.	.	PUNCT
bracis-33576	309	44	:	:	PUNCT
bracis-33576	310	1	language	language	NOUN
bracis-33576	310	2	models	model	NOUN
bracis-33576	310	3	are	be	AUX
bracis-33576	310	4	unsupervised	unsupervised	ADJ
bracis-33576	310	5	multitask	multitask	ADJ
bracis-33576	310	6	learners	learner	NOUN
bracis-33576	310	7	.	.	PUNCT
bracis-33576	311	1	openai	openai	PROPN
bracis-33576	311	2	blog	blog	NOUN
bracis-33576	311	3	1(8	1(8	NUM
bracis-33576	311	4	)	)	PUNCT
bracis-33576	311	5	,	,	PUNCT
bracis-33576	311	6	9	9	NUM
bracis-33576	311	7	(	(	PUNCT
bracis-33576	311	8	2019	2019	NUM
bracis-33576	311	9	)	)	PUNCT
bracis-33576	311	10	google	google	NOUN
bracis-33576	311	11	scholar	scholar	NOUN
bracis-33576	311	12	  	  	SPACE
bracis-33576	311	13	song	song	NOUN
bracis-33576	311	14	,	,	PUNCT
bracis-33576	311	15	k.	k.	PROPN
bracis-33576	311	16	,	,	PUNCT
bracis-33576	311	17	et	et	PROPN
bracis-33576	311	18	al	al	PROPN
bracis-33576	311	19	.	.	PUNCT
bracis-33576	311	20	:	:	PUNCT
bracis-33576	312	1	mpnet	mpnet	NOUN
bracis-33576	312	2	:	:	PUNCT
bracis-33576	312	3	masked	masked	ADJ
bracis-33576	312	4	and	and	CCONJ
bracis-33576	312	5	permuted	permuted	ADJ
bracis-33576	312	6	pre	pre	NOUN
bracis-33576	312	7	-	-	NOUN
bracis-33576	312	8	training	training	NOUN
bracis-33576	312	9	for	for	ADP
bracis-33576	312	10	language	language	NOUN
bracis-33576	312	11	understanding	understanding	NOUN
bracis-33576	312	12	.	.	PUNCT
bracis-33576	313	1	adv	adv	PROPN
bracis-33576	313	2	.	.	PUNCT
bracis-33576	313	3	neural	neural	PROPN
bracis-33576	313	4	inf	inf	PROPN
bracis-33576	313	5	.	.	PUNCT
bracis-33576	313	6	process	process	NOUN
bracis-33576	313	7	.	.	PUNCT
bracis-33576	314	1	syst	syst	PROPN
bracis-33576	314	2	.	.	PUNCT
bracis-33576	315	1	33	33	NUM
bracis-33576	315	2	,	,	PUNCT
bracis-33576	315	3	16857–16867	16857–16867	NUM
bracis-33576	315	4	(	(	PUNCT
bracis-33576	315	5	2020	2020	NUM
bracis-33576	315	6	)	)	PUNCT
bracis-33576	315	7	math	math	NOUN
bracis-33576	315	8	  	  	SPACE
bracis-33576	315	9	google	google	PROPN
bracis-33576	315	10	scholar	scholar	NOUN
bracis-33576	315	11	  	  	SPACE
bracis-33576	315	12	tipirneni	tipirneni	PROPN
bracis-33576	315	13	,	,	PUNCT
bracis-33576	315	14	s.	s.	PROPN
bracis-33576	315	15	,	,	PUNCT
bracis-33576	315	16	adkathimar	adkathimar	PROPN
bracis-33576	315	17	,	,	PUNCT
bracis-33576	315	18	n.	n.	NOUN
bracis-33576	315	19	,	,	PUNCT
bracis-33576	315	20	hiranandani	hiranandani	PROPN
bracis-33576	315	21	,	,	PUNCT
bracis-33576	315	22	v.n	v.n	PROPN
bracis-33576	315	23	.	.	PROPN
bracis-33576	315	24	,	,	PUNCT
bracis-33576	315	25	yuan	yuan	PROPN
bracis-33576	315	26	,	,	PUNCT
bracis-33576	315	27	c.	c.	PROPN
bracis-33576	315	28	,	,	PUNCT
bracis-33576	315	29	reddy	reddy	PROPN
bracis-33576	315	30	,	,	PUNCT
bracis-33576	315	31	c.k	c.k	PROPN
bracis-33576	315	32	.	.	PUNCT
bracis-33576	315	33	:	:	PUNCT
bracis-33576	316	1	context	context	NOUN
bracis-33576	316	2	-	-	PUNCT
bracis-33576	316	3	aware	aware	ADJ
bracis-33576	316	4	clustering	clustering	NOUN
bracis-33576	316	5	using	use	VERB
bracis-33576	316	6	large	large	ADJ
bracis-33576	316	7	language	language	NOUN
bracis-33576	316	8	models	model	NOUN
bracis-33576	316	9	.	.	PUNCT
bracis-33576	317	1	arxiv	arxiv	NOUN
bracis-33576	317	2	(	(	PUNCT
bracis-33576	317	3	2024	2024	NUM
bracis-33576	317	4	)	)	PUNCT
bracis-33576	317	5	google	google	NOUN
bracis-33576	317	6	scholar	scholar	NOUN
bracis-33576	317	7	  	  	SPACE
bracis-33576	317	8	tissera	tissera	ADV
bracis-33576	317	9	,	,	PUNCT
bracis-33576	317	10	m.	m.	NOUN
bracis-33576	317	11	,	,	PUNCT
bracis-33576	317	12	asanka	asanka	NOUN
bracis-33576	317	13	,	,	PUNCT
bracis-33576	317	14	p.	p.	PROPN
bracis-33576	317	15	,	,	PUNCT
bracis-33576	317	16	et	et	PROPN
bracis-33576	317	17	al	al	PROPN
bracis-33576	317	18	.	.	PROPN
bracis-33576	317	19	:	:	PUNCT
bracis-33576	317	20	enhancing	enhance	VERB
bracis-33576	317	21	customer	customer	NOUN
bracis-33576	317	22	segmentation	segmentation	NOUN
bracis-33576	317	23	using	use	VERB
bracis-33576	317	24	llms	llm	NOUN
bracis-33576	317	25	and	and	CCONJ
bracis-33576	317	26	deterministic	deterministic	ADJ
bracis-33576	317	27	,	,	PUNCT
bracis-33576	317	28	independent	independent	ADJ
bracis-33576	317	29	-	-	PUNCT
bracis-33576	317	30	of	of	ADP
bracis-33576	317	31	-	-	PUNCT
bracis-33576	317	32	corpus	corpus	NOUN
bracis-33576	317	33	embeddings	embedding	NOUN
bracis-33576	317	34	.	.	PUNCT
bracis-33576	318	1	in	in	ADP
bracis-33576	318	2	:	:	PUNCT
bracis-33576	318	3	2024	2024	NUM
bracis-33576	318	4	4th	4th	ADJ
bracis-33576	318	5	(	(	PUNCT
bracis-33576	318	6	icarc	icarc	NOUN
bracis-33576	318	7	)	)	PUNCT
bracis-33576	318	8	(	(	PUNCT
bracis-33576	318	9	2024	2024	NUM
bracis-33576	318	10	)	)	PUNCT
bracis-33576	318	11	google	google	NOUN
bracis-33576	318	12	scholar	scholar	NOUN
bracis-33576	318	13	  	  	SPACE
bracis-33576	318	14	touvron	touvron	ADJ
bracis-33576	318	15	,	,	PUNCT
bracis-33576	318	16	h.	h.	PROPN
bracis-33576	318	17	,	,	PUNCT
bracis-33576	318	18	martin	martin	PROPN
bracis-33576	318	19	,	,	PUNCT
bracis-33576	318	20	l.	l.	PROPN
bracis-33576	318	21	,	,	PUNCT
bracis-33576	318	22	stone	stone	PROPN
bracis-33576	318	23	,	,	PUNCT
bracis-33576	318	24	k.	k.	PROPN
bracis-33576	318	25	,	,	PUNCT
bracis-33576	318	26	et	et	PROPN
bracis-33576	318	27	al	al	PROPN
bracis-33576	318	28	.	.	PUNCT
bracis-33576	318	29	:	:	PUNCT
bracis-33576	318	30	llama	llama	PROPN
bracis-33576	318	31	2	2	NUM
bracis-33576	318	32	:	:	PUNCT
bracis-33576	318	33	open	open	ADJ
bracis-33576	318	34	foundation	foundation	NOUN
bracis-33576	318	35	and	and	CCONJ
bracis-33576	318	36	fine	fine	ADV
bracis-33576	318	37	-	-	PUNCT
bracis-33576	318	38	tuned	tune	VERB
bracis-33576	318	39	chat	chat	NOUN
bracis-33576	318	40	models	model	NOUN
bracis-33576	318	41	.	.	PUNCT
bracis-33576	319	1	arxiv	arxiv	PROPN
bracis-33576	319	2	preprint	preprint	VERB
bracis-33576	319	3	arxiv:2307.09288	arxiv:2307.09288	NOUN
bracis-33576	319	4	(	(	PUNCT
bracis-33576	319	5	2023	2023	NUM
bracis-33576	319	6	)	)	PUNCT
bracis-33576	319	7	viswanathan	viswanathan	PROPN
bracis-33576	319	8	,	,	PUNCT
bracis-33576	319	9	v.	v.	ADV
bracis-33576	319	10	,	,	PUNCT
bracis-33576	319	11	gashteovski	gashteovski	NOUN
bracis-33576	319	12	,	,	PUNCT
bracis-33576	319	13	k.	k.	PROPN
bracis-33576	319	14	,	,	PUNCT
bracis-33576	319	15	lawrence	lawrence	PROPN
bracis-33576	319	16	,	,	PUNCT
bracis-33576	319	17	c.	c.	PROPN
bracis-33576	319	18	,	,	PUNCT
bracis-33576	319	19	wu	wu	PROPN
bracis-33576	319	20	,	,	PUNCT
bracis-33576	319	21	t.	t.	PROPN
bracis-33576	319	22	,	,	PUNCT
bracis-33576	319	23	neubig	neubig	ADJ
bracis-33576	319	24	,	,	PUNCT
bracis-33576	319	25	g.	g.	NOUN
bracis-33576	319	26	:	:	PUNCT
bracis-33576	319	27	large	large	ADJ
bracis-33576	319	28	language	language	NOUN
bracis-33576	319	29	models	model	NOUN
bracis-33576	319	30	enable	enable	VERB
bracis-33576	319	31	few	few	ADJ
bracis-33576	319	32	-	-	PUNCT
bracis-33576	319	33	shot	shot	NOUN
bracis-33576	319	34	clustering	clustering	NOUN
bracis-33576	319	35	.	.	PUNCT
bracis-33576	320	1	arxiv	arxiv	PROPN
bracis-33576	320	2	preprint	preprint	VERB
bracis-33576	320	3	arxiv:2307.00524	arxiv:2307.00524	NUM
bracis-33576	320	4	(	(	PUNCT
bracis-33576	320	5	2023	2023	NUM
bracis-33576	320	6	)	)	PUNCT
bracis-33576	320	7	warner	warner	PROPN
bracis-33576	320	8	,	,	PUNCT
bracis-33576	320	9	j.	j.	PROPN
bracis-33576	320	10	,	,	PUNCT
bracis-33576	320	11	sexauer	sexauer	PROPN
bracis-33576	320	12	,	,	PUNCT
bracis-33576	320	13	j.	j.	PROPN
bracis-33576	320	14	,	,	PUNCT
bracis-33576	320	15	et	et	PROPN
bracis-33576	320	16	 	 	SPACE
bracis-33576	320	17	al	al	PROPN
bracis-33576	320	18	.	.	PROPN
bracis-33576	320	19	:	:	PUNCT
bracis-33576	321	1	jdwarner	jdwarner	NOUN
bracis-33576	321	2	/	/	SYM
bracis-33576	321	3	scikit	scikit	NOUN
bracis-33576	321	4	-	-	PUNCT
bracis-33576	321	5	fuzzy	fuzzy	ADJ
bracis-33576	321	6	:	:	PUNCT
bracis-33576	321	7	scikit	scikit	NOUN
bracis-33576	321	8	-	-	PUNCT
bracis-33576	321	9	fuzzy	fuzzy	ADJ
bracis-33576	321	10	version	version	NOUN
bracis-33576	321	11	0.4.2	0.4.2	NOUN
bracis-33576	321	12	.	.	PUNCT
bracis-33576	322	1	zenodo	zenodo	NOUN
bracis-33576	322	2	(	(	PUNCT
bracis-33576	322	3	2019	2019	NUM
bracis-33576	322	4	)	)	PUNCT
bracis-33576	322	5	.	.	PUNCT
bracis-33576	323	1	https://doi.org/10.5281/zenodo.3541386	https://doi.org/10.5281/zenodo.3541386	PROPN
bracis-33576	323	2	zhang	zhang	PROPN
bracis-33576	323	3	,	,	PUNCT
bracis-33576	323	4	y.	y.	PROPN
bracis-33576	323	5	,	,	PUNCT
bracis-33576	323	6	wang	wang	PROPN
bracis-33576	323	7	,	,	PUNCT
bracis-33576	323	8	z.	z.	PROPN
bracis-33576	323	9	,	,	PUNCT
bracis-33576	323	10	shang	shang	PROPN
bracis-33576	323	11	,	,	PUNCT
bracis-33576	323	12	j.	j.	PROPN
bracis-33576	323	13	:	:	PUNCT
bracis-33576	323	14	clusterllm	clusterllm	NOUN
bracis-33576	323	15	:	:	PUNCT
bracis-33576	323	16	large	large	ADJ
bracis-33576	323	17	language	language	NOUN
bracis-33576	323	18	models	model	NOUN
bracis-33576	323	19	as	as	ADP
bracis-33576	323	20	a	a	DET
bracis-33576	323	21	guide	guide	NOUN
bracis-33576	323	22	for	for	ADP
bracis-33576	323	23	text	text	NOUN
bracis-33576	323	24	clustering	clustering	NOUN
bracis-33576	323	25	.	.	PUNCT
bracis-33576	324	1	arxiv	arxiv	PROPN
bracis-33576	324	2	preprint	preprint	NOUN
bracis-33576	324	3	arxiv:2305.14871	arxiv:2305.14871	PROPN
bracis-33576	324	4	(	(	PUNCT
bracis-33576	324	5	2023	2023	NUM
bracis-33576	324	6	)	)	PUNCT
bracis-33576	324	7	zhu	zhu	PROPN
bracis-33576	324	8	,	,	PUNCT
bracis-33576	324	9	p.	p.	PROPN
bracis-33576	324	10	,	,	PUNCT
bracis-33576	324	11	lang	lang	PROPN
bracis-33576	324	12	,	,	PUNCT
bracis-33576	324	13	q.	q.	PROPN
bracis-33576	324	14	,	,	PUNCT
bracis-33576	324	15	liu	liu	PROPN
bracis-33576	324	16	,	,	PUNCT
bracis-33576	324	17	x.	x.	NOUN
bracis-33576	324	18	:	:	PUNCT
bracis-33576	324	19	word	word	NOUN
bracis-33576	324	20	embedding	embed	VERB
bracis-33576	324	21	of	of	ADP
bracis-33576	324	22	dimensionality	dimensionality	NOUN
bracis-33576	324	23	reduction	reduction	NOUN
bracis-33576	324	24	for	for	ADP
bracis-33576	324	25	document	document	NOUN
bracis-33576	324	26	clustering	clustering	NOUN
bracis-33576	324	27	.	.	PUNCT
bracis-33576	325	1	in	in	ADP
bracis-33576	325	2	:	:	PUNCT
bracis-33576	325	3	2023	2023	NUM
bracis-33576	325	4	35th	35th	ADJ
bracis-33576	325	5	chinese	chinese	ADJ
bracis-33576	325	6	control	control	NOUN
bracis-33576	325	7	and	and	CCONJ
bracis-33576	325	8	decision	decision	NOUN
bracis-33576	325	9	conference	conference	NOUN
bracis-33576	325	10	(	(	PUNCT
bracis-33576	325	11	ccdc	ccdc	PROPN
bracis-33576	325	12	)	)	PUNCT
bracis-33576	325	13	,	,	PUNCT
bracis-33576	325	14	pp	pp	PROPN
bracis-33576	325	15	.	.	PUNCT
bracis-33576	325	16	4371–4376	4371–4376	X
bracis-33576	325	17	.	.	PUNCT
bracis-33576	326	1	ieee	ieee	NOUN
bracis-33576	326	2	(	(	PUNCT
bracis-33576	326	3	2023	2023	NUM
bracis-33576	326	4	)	)	PUNCT
bracis-33576	326	5	google	google	NOUN
bracis-33576	326	6	scholar	scholar	NOUN
bracis-33576	326	7	  	  	SPACE
bracis-33576	326	8	download	download	NOUN
bracis-33576	326	9	references	reference	NOUN
bracis-33576	326	10	acknowledgements	acknowledgement	NOUN
bracis-33576	326	11	this	this	DET
bracis-33576	326	12	project	project	NOUN
bracis-33576	326	13	was	be	AUX
bracis-33576	326	14	supported	support	VERB
bracis-33576	326	15	by	by	ADP
bracis-33576	326	16	the	the	DET
bracis-33576	326	17	brazilian	brazilian	PROPN
bracis-33576	326	18	ministry	ministry	PROPN
bracis-33576	326	19	of	of	ADP
bracis-33576	326	20	science	science	PROPN
bracis-33576	326	21	,	,	PUNCT
bracis-33576	326	22	technology	technology	NOUN
bracis-33576	326	23	and	and	CCONJ
bracis-33576	326	24	innovations	innovation	NOUN
bracis-33576	326	25	,	,	PUNCT
bracis-33576	326	26	with	with	ADP
bracis-33576	326	27	resources	resource	NOUN
bracis-33576	326	28	from	from	ADP
bracis-33576	326	29	law	law	NOUN
bracis-33576	326	30	n\(^{\underline{{\textbf	n\(^{\underline{{\textbf	NOUN
bracis-33576	326	31	{	{	PUNCT
bracis-33576	326	32	o}}}}\	o}}}}\	PROPN
bracis-33576	326	33	)	)	PUNCT
bracis-33576	326	34	8,248	8,248	NUM
bracis-33576	326	35	,	,	PUNCT
bracis-33576	326	36	of	of	ADP
bracis-33576	326	37	october	october	PROPN
bracis-33576	326	38	23	23	NUM
bracis-33576	326	39	,	,	PUNCT
bracis-33576	326	40	1991	1991	NUM
bracis-33576	326	41	,	,	PUNCT
bracis-33576	326	42	within	within	ADP
bracis-33576	326	43	the	the	DET
bracis-33576	326	44	scope	scope	NOUN
bracis-33576	326	45	of	of	ADP
bracis-33576	326	46	ppi	ppi	NOUN
bracis-33576	326	47	-	-	PUNCT
bracis-33576	326	48	softex	softex	NOUN
bracis-33576	326	49	,	,	PUNCT
bracis-33576	326	50	coordinated	coordinate	VERB
bracis-33576	326	51	by	by	ADP
bracis-33576	326	52	softex	softex	NOUN
bracis-33576	326	53	and	and	CCONJ
bracis-33576	326	54	published	publish	VERB
bracis-33576	326	55	arquitetura	arquitetura	PROPN
bracis-33576	326	56	cognitiva	cognitiva	NOUN
bracis-33576	326	57	(	(	PUNCT
bracis-33576	326	58	phase	phase	NOUN
bracis-33576	326	59	3	3	NUM
bracis-33576	326	60	)	)	PUNCT
bracis-33576	326	61	,	,	PUNCT
bracis-33576	326	62	dou	dou	X
bracis-33576	326	63	01245.003479/2024	01245.003479/2024	ADJ
bracis-33576	326	64	-10	-10	PROPN
bracis-33576	326	65	.	.	PUNCT
bracis-33576	327	1	author	author	NOUN
bracis-33576	327	2	information	information	NOUN
bracis-33576	327	3	authors	author	NOUN
bracis-33576	327	4	and	and	CCONJ
bracis-33576	327	5	affiliations	affiliation	NOUN
bracis-33576	327	6	hub	hub	PROPN
bracis-33576	327	7	de	de	PROPN
bracis-33576	327	8	inteligência	inteligência	PROPN
bracis-33576	327	9	artificial	artificial	ADJ
bracis-33576	327	10	e	e	NOUN
bracis-33576	327	11	arquiteturas	arquitetura	VERB
bracis-33576	327	12	cognitivas	cognitivas	PROPN
bracis-33576	327	13	(	(	PUNCT
bracis-33576	327	14	h.iaac	h.iaac	NOUN
bracis-33576	327	15	)	)	PUNCT
bracis-33576	327	16	,	,	PUNCT
bracis-33576	327	17	campinas	campinas	PROPN
bracis-33576	327	18	,	,	PUNCT
bracis-33576	327	19	brazil	brazil	PROPN
bracis-33576	327	20	fillipe	fillipe	PROPN
bracis-33576	327	21	dos	dos	PROPN
bracis-33576	327	22	santos	santos	PROPN
bracis-33576	327	23	silva	silva	PROPN
bracis-33576	327	24	,	,	PUNCT
bracis-33576	327	25	 	 	SPACE
bracis-33576	327	26	gabriel	gabriel	PROPN
bracis-33576	327	27	kenzo	kenzo	PROPN
bracis-33576	327	28	kakimoto	kakimoto	PROPN
bracis-33576	327	29	,	,	PUNCT
bracis-33576	327	30	 	 	SPACE
bracis-33576	327	31	julio	julio	PROPN
bracis-33576	327	32	cesar	cesar	PROPN
bracis-33576	327	33	dos	dos	PROPN
bracis-33576	327	34	reis	reis	NOUN
bracis-33576	327	35	 	 	SPACE
bracis-33576	327	36	&	&	CCONJ
bracis-33576	327	37	 	 	SPACE
bracis-33576	327	38	marcelo	marcelo	PROPN
bracis-33576	327	39	s.	s.	PROPN
bracis-33576	327	40	reis	reis	VERB
bracis-33576	327	41	artificial	artificial	ADJ
bracis-33576	327	42	intelligence	intelligence	NOUN
bracis-33576	327	43	laboratory	laboratory	NOUN
bracis-33576	327	44	(	(	PUNCT
bracis-33576	327	45	recod.ai	recod.ai	PROPN
bracis-33576	327	46	)	)	PUNCT
bracis-33576	327	47	,	,	PUNCT
bracis-33576	327	48	campinas	campinas	PROPN
bracis-33576	327	49	,	,	PUNCT
bracis-33576	327	50	brazil	brazil	PROPN
bracis-33576	327	51	fillipe	fillipe	PROPN
bracis-33576	327	52	dos	dos	PROPN
bracis-33576	327	53	santos	santos	PROPN
bracis-33576	327	54	silva	silva	PROPN
bracis-33576	327	55	,	,	PUNCT
bracis-33576	327	56	 	 	SPACE
bracis-33576	327	57	gabriel	gabriel	PROPN
bracis-33576	327	58	kenzo	kenzo	PROPN
bracis-33576	327	59	kakimoto	kakimoto	PROPN
bracis-33576	327	60	 	 	SPACE
bracis-33576	327	61	&	&	CCONJ
bracis-33576	327	62	 	 	SPACE
bracis-33576	327	63	marcelo	marcelo	PROPN
bracis-33576	327	64	s.	s.	PROPN
bracis-33576	327	65	reis	reis	PROPN
bracis-33576	327	66	instituto	instituto	PROPN
bracis-33576	327	67	de	de	PROPN
bracis-33576	327	68	computação	computação	PROPN
bracis-33576	327	69	,	,	PUNCT
bracis-33576	327	70	universidade	universidade	PROPN
bracis-33576	327	71	estadual	estadual	PROPN
bracis-33576	327	72	de	de	PROPN
bracis-33576	327	73	campinas	campinas	PROPN
bracis-33576	327	74	(	(	PUNCT
bracis-33576	327	75	unicamp	unicamp	PROPN
bracis-33576	327	76	)	)	PUNCT
bracis-33576	327	77	,	,	PUNCT
bracis-33576	327	78	campinas	campinas	PROPN
bracis-33576	327	79	,	,	PUNCT
bracis-33576	327	80	brazil	brazil	PROPN
bracis-33576	327	81	fillipe	fillipe	PROPN
bracis-33576	327	82	dos	dos	PROPN
bracis-33576	327	83	santos	santos	PROPN
bracis-33576	327	84	silva	silva	PROPN
bracis-33576	327	85	,	,	PUNCT
bracis-33576	327	86	 	 	SPACE
bracis-33576	327	87	julio	julio	PROPN
bracis-33576	327	88	cesar	cesar	PROPN
bracis-33576	327	89	dos	dos	PROPN
bracis-33576	327	90	reis	reis	NOUN
bracis-33576	327	91	 	 	SPACE
bracis-33576	327	92	&	&	CCONJ
bracis-33576	327	93	 	 	SPACE
bracis-33576	327	94	marcelo	marcelo	PROPN
bracis-33576	327	95	s.	s.	PROPN
bracis-33576	327	96	reis	reis	PROPN
bracis-33576	327	97	faculdade	faculdade	PROPN
bracis-33576	327	98	de	de	PROPN
bracis-33576	327	99	engenharia	engenharia	PROPN
bracis-33576	327	100	mecânica	mecânica	PROPN
bracis-33576	327	101	,	,	PUNCT
bracis-33576	327	102	universidade	universidade	PROPN
bracis-33576	327	103	estadual	estadual	PROPN
bracis-33576	327	104	de	de	PROPN
bracis-33576	327	105	campinas	campinas	PROPN
bracis-33576	327	106	(	(	PUNCT
bracis-33576	327	107	unicamp	unicamp	PROPN
bracis-33576	327	108	)	)	PUNCT
bracis-33576	327	109	,	,	PUNCT
bracis-33576	327	110	campinas	campinas	PROPN
bracis-33576	327	111	,	,	PUNCT
bracis-33576	327	112	brazil	brazil	PROPN
bracis-33576	327	113	gabriel	gabriel	PROPN
bracis-33576	327	114	kenzo	kenzo	PROPN
bracis-33576	327	115	kakimoto	kakimoto	PROPN
bracis-33576	327	116	authors	author	NOUN
bracis-33576	327	117	fillipe	fillipe	VERB
bracis-33576	327	118	dos	dos	PROPN
bracis-33576	327	119	santos	santo	NOUN
bracis-33576	327	120	silvaview	silvaview	VERB
bracis-33576	327	121	author	author	NOUN
bracis-33576	327	122	publications	publication	NOUN
bracis-33576	327	123	search	search	NOUN
bracis-33576	327	124	author	author	NOUN
bracis-33576	327	125	on	on	ADP
bracis-33576	327	126	:	:	PUNCT
bracis-33576	327	127	pubmed	pubmed	PROPN
bracis-33576	327	128	 	 	SPACE
bracis-33576	327	129	google	google	PROPN
bracis-33576	327	130	scholar	scholar	PROPN
bracis-33576	327	131	gabriel	gabriel	PROPN
bracis-33576	327	132	kenzo	kenzo	PROPN
bracis-33576	327	133	kakimotoview	kakimotoview	PROPN
bracis-33576	327	134	author	author	NOUN
bracis-33576	327	135	publications	publication	NOUN
bracis-33576	327	136	search	search	NOUN
bracis-33576	327	137	author	author	NOUN
bracis-33576	327	138	on	on	ADP
bracis-33576	327	139	:	:	PUNCT
bracis-33576	327	140	pubmed	pubmed	PROPN
bracis-33576	327	141	 	 	SPACE
bracis-33576	327	142	google	google	PROPN
bracis-33576	327	143	scholar	scholar	PROPN
bracis-33576	327	144	julio	julio	PROPN
bracis-33576	327	145	cesar	cesar	PROPN
bracis-33576	327	146	dos	dos	AUX
bracis-33576	327	147	reisview	reisview	VERB
bracis-33576	327	148	author	author	NOUN
bracis-33576	327	149	publications	publication	NOUN
bracis-33576	327	150	search	search	NOUN
bracis-33576	327	151	author	author	NOUN
bracis-33576	327	152	on	on	ADP
bracis-33576	327	153	:	:	PUNCT
bracis-33576	327	154	pubmed	pubmed	PROPN
bracis-33576	327	155	 	 	SPACE
bracis-33576	327	156	google	google	PROPN
bracis-33576	327	157	scholar	scholar	PROPN
bracis-33576	327	158	marcelo	marcelo	PROPN
bracis-33576	327	159	s.	s.	PROPN
bracis-33576	327	160	reisview	reisview	VERB
bracis-33576	327	161	author	author	NOUN
bracis-33576	327	162	publications	publication	NOUN
bracis-33576	327	163	search	search	NOUN
bracis-33576	327	164	author	author	NOUN
bracis-33576	327	165	on	on	ADP
bracis-33576	327	166	:	:	PUNCT
bracis-33576	327	167	pubmed	pubmed	PROPN
bracis-33576	327	168	 	 	SPACE
bracis-33576	327	169	google	google	PROPN
bracis-33576	327	170	scholar	scholar	NOUN
bracis-33576	327	171	corresponding	correspond	VERB
bracis-33576	327	172	author	author	NOUN
bracis-33576	327	173	correspondence	correspondence	NOUN
bracis-33576	327	174	to	to	ADP
bracis-33576	327	175	fillipe	fillipe	VERB
bracis-33576	327	176	dos	dos	PROPN
bracis-33576	327	177	santos	santos	PROPN
bracis-33576	327	178	silva	silva	PROPN
bracis-33576	327	179	.	.	PUNCT
bracis-33576	328	1	editor	editor	NOUN
bracis-33576	328	2	information	information	NOUN
bracis-33576	328	3	editors	editor	NOUN
bracis-33576	328	4	and	and	CCONJ
bracis-33576	328	5	affiliations	affiliation	NOUN
bracis-33576	328	6	universidade	universidade	VERB
bracis-33576	328	7	federal	federal	PROPN
bracis-33576	328	8	fluminense	fluminense	PROPN
bracis-33576	328	9	,	,	PUNCT
bracis-33576	328	10	niterói	niterói	PROPN
bracis-33576	328	11	,	,	PUNCT
bracis-33576	328	12	brazil	brazil	PROPN
bracis-33576	328	13	aline	aline	PROPN
bracis-33576	328	14	paes	paes	PROPN
bracis-33576	328	15	instituto	instituto	PROPN
bracis-33576	328	16	tecnológico	tecnológico	PROPN
bracis-33576	328	17	de	de	PROPN
bracis-33576	328	18	aeronáutica	aeronáutica	PROPN
bracis-33576	328	19	,	,	PUNCT
bracis-33576	328	20	são	são	PROPN
bracis-33576	328	21	josé	josé	PROPN
bracis-33576	328	22	dos	dos	PROPN
bracis-33576	328	23	campos	campos	PROPN
bracis-33576	328	24	,	,	PUNCT
bracis-33576	328	25	brazil	brazil	PROPN
bracis-33576	328	26	filipe	filipe	PROPN
bracis-33576	328	27	a.	a.	PROPN
bracis-33576	328	28	n.	n.	PROPN
bracis-33576	328	29	verri	verri	PROPN
bracis-33576	328	30	rights	right	NOUN
bracis-33576	328	31	and	and	CCONJ
bracis-33576	328	32	permissions	permission	NOUN
bracis-33576	328	33	reprints	reprint	NOUN
bracis-33576	328	34	and	and	CCONJ
bracis-33576	328	35	permissions	permission	VERB
bracis-33576	328	36	copyright	copyright	NOUN
bracis-33576	328	37	information	information	NOUN
bracis-33576	328	38	©	©	ADP
bracis-33576	328	39	2025	2025	NUM
bracis-33576	328	40	the	the	DET
bracis-33576	328	41	author(s	author(s	NOUN
bracis-33576	328	42	)	)	PUNCT
bracis-33576	328	43	,	,	PUNCT
bracis-33576	328	44	under	under	ADP
bracis-33576	328	45	exclusive	exclusive	ADJ
bracis-33576	328	46	license	license	NOUN
bracis-33576	328	47	to	to	ADP
bracis-33576	328	48	springer	springer	NOUN
bracis-33576	328	49	nature	nature	PROPN
bracis-33576	328	50	switzerland	switzerland	PROPN
bracis-33576	328	51	ag	ag	PROPN
bracis-33576	328	52	about	about	ADP
bracis-33576	328	53	this	this	DET
bracis-33576	328	54	paper	paper	NOUN
bracis-33576	328	55	cite	cite	VERB
bracis-33576	328	56	this	this	DET
bracis-33576	328	57	paper	paper	NOUN
bracis-33576	328	58	dos	do	NOUN
bracis-33576	328	59	santos	santos	PROPN
bracis-33576	328	60	silva	silva	PROPN
bracis-33576	328	61	,	,	PUNCT
bracis-33576	328	62	f.	f.	PROPN
bracis-33576	328	63	,	,	PUNCT
bracis-33576	328	64	kakimoto	kakimoto	PROPN
bracis-33576	328	65	,	,	PUNCT
bracis-33576	328	66	g.k	g.k	PROPN
bracis-33576	328	67	.	.	PROPN
bracis-33576	328	68	,	,	PUNCT
bracis-33576	328	69	dos	do	VERB
bracis-33576	328	70	reis	reis	NOUN
bracis-33576	328	71	,	,	PUNCT
bracis-33576	328	72	j.c	j.c	PROPN
bracis-33576	328	73	.	.	PROPN
bracis-33576	328	74	,	,	PUNCT
bracis-33576	328	75	reis	reis	NOUN
bracis-33576	328	76	,	,	PUNCT
bracis-33576	328	77	m.s	m.s	PROPN
bracis-33576	328	78	.	.	PROPN
bracis-33576	328	79	(	(	PUNCT
bracis-33576	328	80	2025	2025	NUM
bracis-33576	328	81	)	)	PUNCT
bracis-33576	328	82	.	.	PUNCT
bracis-33576	329	1	erasmo	erasmo	PROPN
bracis-33576	329	2	:	:	PUNCT
bracis-33576	329	3	leveraging	leverage	VERB
bracis-33576	329	4	large	large	ADJ
bracis-33576	329	5	language	language	NOUN
bracis-33576	329	6	models	model	NOUN
bracis-33576	329	7	for	for	ADP
bracis-33576	329	8	 	 	SPACE
bracis-33576	329	9	enhanced	enhance	VERB
bracis-33576	329	10	clustering	clustering	ADJ
bracis-33576	329	11	segmentation	segmentation	NOUN
bracis-33576	329	12	.	.	PUNCT
bracis-33576	330	1	in	in	ADP
bracis-33576	330	2	:	:	PUNCT
bracis-33576	330	3	paes	paes	PROPN
bracis-33576	330	4	,	,	PUNCT
bracis-33576	330	5	a.	a.	PROPN
bracis-33576	330	6	,	,	PUNCT
bracis-33576	330	7	verri	verri	NOUN
bracis-33576	330	8	,	,	PUNCT
bracis-33576	330	9	f.a.n	f.a.n	ADJ
bracis-33576	330	10	.	.	PUNCT
bracis-33576	330	11	(	(	PUNCT
bracis-33576	330	12	eds	ed	NOUN
bracis-33576	330	13	)	)	PUNCT
bracis-33576	330	14	intelligent	intelligent	ADJ
bracis-33576	330	15	systems	system	NOUN
bracis-33576	330	16	.	.	PUNCT
bracis-33576	330	17	bracis	bracis	PROPN
bracis-33576	330	18	2024	2024	NUM
bracis-33576	330	19	.	.	PUNCT
bracis-33576	331	1	lecture	lecture	NOUN
bracis-33576	331	2	notes	note	NOUN
bracis-33576	331	3	in	in	ADP
bracis-33576	331	4	computer	computer	NOUN
bracis-33576	331	5	science	science	NOUN
bracis-33576	331	6	(	(	PUNCT
bracis-33576	331	7	)	)	PUNCT
bracis-33576	331	8	,	,	PUNCT
bracis-33576	331	9	vol	vol	NOUN
bracis-33576	331	10	15412	15412	NUM
bracis-33576	331	11	.	.	PUNCT
bracis-33576	332	1	springer	springer	NOUN
bracis-33576	332	2	,	,	PUNCT
bracis-33576	332	3	cham	cham	PROPN
bracis-33576	332	4	.	.	PUNCT
bracis-33576	333	1	https://doi.org/10.1007/978-3-031-79029-4_29	https://doi.org/10.1007/978-3-031-79029-4_29	PROPN
bracis-33576	333	2	download	download	NOUN
bracis-33576	333	3	citation	citation	NOUN
bracis-33576	333	4	.ris	.ris	PUNCT
bracis-33576	334	1	.enw	.enw	PROPN
bracis-33576	334	2	.bib	.bib	PUNCT
bracis-33576	335	1	doi	doi	PROPN
bracis-33576	335	2	:	:	PUNCT
bracis-33576	335	3	https://doi.org/10.1007/978-3-031-79029-4_29	https://doi.org/10.1007/978-3-031-79029-4_29	PROPN
bracis-33576	335	4	published	publish	VERB
bracis-33576	335	5	:	:	PUNCT
bracis-33576	335	6	30	30	NUM
bracis-33576	335	7	january	january	PROPN
bracis-33576	335	8	2025	2025	NUM
bracis-33576	335	9	publisher	publisher	NOUN
bracis-33576	335	10	name	name	NOUN
bracis-33576	335	11	:	:	PUNCT
bracis-33576	335	12	springer	springer	NOUN
bracis-33576	335	13	,	,	PUNCT
bracis-33576	335	14	cham	cham	PROPN
bracis-33576	335	15	print	print	PROPN
bracis-33576	335	16	isbn	isbn	PROPN
bracis-33576	335	17	:	:	PUNCT
bracis-33576	335	18	978	978	NUM
bracis-33576	335	19	-	-	SYM
bracis-33576	335	20	3	3	NUM
bracis-33576	335	21	-	-	PUNCT
bracis-33576	335	22	031	031	NUM
bracis-33576	335	23	-	-	PUNCT
bracis-33576	335	24	79028	79028	NUM
bracis-33576	335	25	-	-	SYM
bracis-33576	335	26	7	7	NUM
bracis-33576	335	27	online	online	ADJ
bracis-33576	335	28	isbn	isbn	NOUN
bracis-33576	335	29	:	:	PUNCT
bracis-33576	335	30	978	978	NUM
bracis-33576	335	31	-	-	SYM
bracis-33576	335	32	3	3	NUM
bracis-33576	335	33	-	-	PUNCT
bracis-33576	335	34	031	031	NUM
bracis-33576	335	35	-	-	PUNCT
bracis-33576	335	36	79029	79029	NUM
bracis-33576	335	37	-	-	SYM
bracis-33576	335	38	4	4	NUM
bracis-33576	335	39	ebook	ebook	NOUN
bracis-33576	335	40	packages	package	NOUN
bracis-33576	335	41	:	:	PUNCT
bracis-33576	335	42	computer	computer	NOUN
bracis-33576	335	43	sciencecomputer	sciencecomputer	NOUN
bracis-33576	335	44	science	science	NOUN
bracis-33576	335	45	(	(	PUNCT
bracis-33576	335	46	r0	r0	NOUN
bracis-33576	335	47	)	)	PUNCT
bracis-33576	335	48	share	share	VERB
bracis-33576	335	49	this	this	DET
bracis-33576	335	50	paper	paper	NOUN
bracis-33576	335	51	anyone	anyone	PRON
bracis-33576	335	52	you	you	PRON
bracis-33576	335	53	share	share	VERB
bracis-33576	335	54	the	the	DET
bracis-33576	335	55	following	follow	VERB
bracis-33576	335	56	link	link	NOUN
bracis-33576	335	57	with	with	ADP
bracis-33576	335	58	will	will	AUX
bracis-33576	335	59	be	be	AUX
bracis-33576	335	60	able	able	ADJ
bracis-33576	335	61	to	to	PART
bracis-33576	335	62	read	read	VERB
bracis-33576	335	63	this	this	DET
bracis-33576	335	64	content	content	NOUN
bracis-33576	335	65	:	:	PUNCT
bracis-33576	335	66	get	get	VERB
bracis-33576	335	67	shareable	shareable	ADJ
bracis-33576	335	68	linksorry	linksorry	NOUN
bracis-33576	335	69	,	,	PUNCT
bracis-33576	335	70	a	a	DET
bracis-33576	335	71	shareable	shareable	ADJ
bracis-33576	335	72	link	link	NOUN
bracis-33576	335	73	is	be	AUX
bracis-33576	335	74	not	not	PART
bracis-33576	335	75	currently	currently	ADV
bracis-33576	335	76	available	available	ADJ
bracis-33576	335	77	for	for	ADP
bracis-33576	335	78	this	this	DET
bracis-33576	335	79	article	article	NOUN
bracis-33576	335	80	.	.	PUNCT
bracis-33576	336	1	copy	copy	VERB
bracis-33576	336	2	shareable	shareable	ADJ
bracis-33576	336	3	link	link	NOUN
bracis-33576	336	4	to	to	PART
bracis-33576	336	5	clipboard	clipboard	NOUN
bracis-33576	336	6	provided	provide	VERB
bracis-33576	336	7	by	by	ADP
bracis-33576	336	8	the	the	DET
bracis-33576	336	9	springer	springer	NOUN
bracis-33576	336	10	nature	nature	PROPN
bracis-33576	336	11	sharedit	sharedit	PROPN
bracis-33576	336	12	content	content	NOUN
bracis-33576	336	13	-	-	PUNCT
bracis-33576	336	14	sharing	share	VERB
bracis-33576	336	15	initiative	initiative	NOUN
bracis-33576	336	16	keywords	keyword	NOUN
bracis-33576	336	17	clustering	cluster	VERB
bracis-33576	336	18	segmentation	segmentation	NOUN
bracis-33576	336	19	transformer	transformer	NOUN
bracis-33576	336	20	-	-	PUNCT
bracis-33576	336	21	based	base	VERB
bracis-33576	336	22	models	model	NOUN
bracis-33576	336	23	tabular	tabular	VERB
bracis-33576	336	24	data	datum	NOUN
bracis-33576	336	25	embeddings	embedding	NOUN
bracis-33576	336	26	publish	publish	VERB
bracis-33576	336	27	with	with	ADP
bracis-33576	336	28	us	us	PROPN
bracis-33576	336	29	policies	policy	NOUN
bracis-33576	336	30	and	and	CCONJ
bracis-33576	336	31	ethics	ethic	NOUN
bracis-33576	336	32	search	search	NOUN
bracis-33576	336	33	search	search	NOUN
bracis-33576	336	34	by	by	ADP
bracis-33576	336	35	keyword	keyword	NOUN
bracis-33576	336	36	or	or	CCONJ
bracis-33576	336	37	author	author	NOUN
bracis-33576	336	38	search	search	NOUN
bracis-33576	336	39	navigation	navigation	NOUN
bracis-33576	336	40	find	find	VERB
bracis-33576	336	41	a	a	DET
bracis-33576	336	42	journal	journal	NOUN
bracis-33576	336	43	publish	publish	VERB
bracis-33576	336	44	with	with	ADP
bracis-33576	336	45	us	we	PRON
bracis-33576	336	46	track	track	VERB
bracis-33576	336	47	your	your	PRON
bracis-33576	336	48	research	research	NOUN
bracis-33576	336	49	discover	discover	VERB
bracis-33576	336	50	content	content	NOUN
bracis-33576	336	51	journals	journal	NOUN
bracis-33576	336	52	a	a	DET
bracis-33576	336	53	-	-	PUNCT
bracis-33576	336	54	z	z	NOUN
bracis-33576	336	55	books	book	NOUN
bracis-33576	336	56	a	a	DET
bracis-33576	336	57	-	-	PUNCT
bracis-33576	336	58	z	z	NOUN
bracis-33576	336	59	publish	publish	NOUN
bracis-33576	336	60	with	with	ADP
bracis-33576	336	61	us	us	PROPN
bracis-33576	336	62	journal	journal	PROPN
bracis-33576	336	63	finder	finder	PROPN
bracis-33576	336	64	publish	publish	VERB
bracis-33576	336	65	your	your	PRON
bracis-33576	336	66	research	research	NOUN
bracis-33576	336	67	language	language	NOUN
bracis-33576	336	68	editing	edit	VERB
bracis-33576	336	69	open	open	ADJ
bracis-33576	336	70	access	access	NOUN
bracis-33576	336	71	publishing	publishing	NOUN
bracis-33576	336	72	products	product	NOUN
bracis-33576	336	73	and	and	CCONJ
bracis-33576	336	74	services	service	NOUN
bracis-33576	336	75	our	our	PRON
bracis-33576	336	76	products	product	NOUN
bracis-33576	336	77	librarians	librarian	VERB
bracis-33576	336	78	societies	society	NOUN
bracis-33576	336	79	partners	partner	NOUN
bracis-33576	336	80	and	and	CCONJ
bracis-33576	336	81	advertisers	advertiser	NOUN
bracis-33576	336	82	our	our	PRON
bracis-33576	336	83	brands	brand	NOUN
bracis-33576	336	84	springer	springer	NOUN
bracis-33576	336	85	nature	nature	PROPN
bracis-33576	336	86	portfolio	portfolio	PROPN
bracis-33576	336	87	bmc	bmc	PROPN
bracis-33576	336	88	palgrave	palgrave	PROPN
bracis-33576	336	89	macmillan	macmillan	PROPN
bracis-33576	336	90	apress	apress	PROPN
bracis-33576	336	91	discover	discover	VERB
bracis-33576	336	92	your	your	PRON
bracis-33576	336	93	privacy	privacy	NOUN
bracis-33576	336	94	choices	choice	NOUN
bracis-33576	336	95	/	/	SYM
bracis-33576	336	96	manage	manage	NOUN
bracis-33576	336	97	cookies	cookie	NOUN
bracis-33576	336	98	your	your	PRON
bracis-33576	336	99	us	us	PROPN
bracis-33576	337	1	state	state	NOUN
bracis-33576	337	2	privacy	privacy	NOUN
bracis-33576	337	3	rights	right	NOUN
bracis-33576	337	4	accessibility	accessibility	NOUN
bracis-33576	337	5	statement	statement	NOUN
bracis-33576	337	6	terms	term	NOUN
bracis-33576	337	7	and	and	CCONJ
bracis-33576	337	8	conditions	condition	NOUN
bracis-33576	337	9	privacy	privacy	NOUN
bracis-33576	337	10	policy	policy	NOUN
bracis-33576	337	11	help	help	NOUN
bracis-33576	337	12	and	and	CCONJ
bracis-33576	337	13	support	support	VERB
bracis-33576	337	14	legal	legal	ADJ
bracis-33576	337	15	notice	notice	NOUN
bracis-33576	337	16	cancel	cancel	VERB
bracis-33576	337	17	contracts	contract	NOUN
bracis-33576	337	18	here	here	ADV
bracis-33576	337	19	129.74.145.123	129.74.145.123	NUM
bracis-33576	337	20	hesburgh	hesburgh	PROPN
bracis-33576	337	21	library	library	PROPN
bracis-33576	337	22	er	er	INTJ
bracis-33576	337	23	unit	unit	NOUN
bracis-33576	337	24	(	(	PUNCT
bracis-33576	337	25	3005732405	3005732405	NUM
bracis-33576	337	26	)	)	PUNCT
bracis-33576	337	27	northeast	northeast	ADJ
bracis-33576	337	28	research	research	NOUN
bracis-33576	337	29	libraries	library	NOUN
bracis-33576	337	30	(	(	PUNCT
bracis-33576	337	31	nerl	nerl	PROPN
bracis-33576	337	32	)	)	PUNCT
bracis-33576	337	33	(	(	PUNCT
bracis-33576	337	34	8200828607	8200828607	NUM
bracis-33576	337	35	)	)	PUNCT
bracis-33576	337	36	nerl	nerl	VERB
bracis-33576	337	37	ta	ta	X
bracis-33576	337	38	account	account	NOUN
bracis-33576	337	39	(	(	PUNCT
bracis-33576	337	40	3006206169	3006206169	NUM
bracis-33576	337	41	)	)	PUNCT
bracis-33576	337	42	university	university	NOUN
bracis-33576	337	43	of	of	ADP
bracis-33576	337	44	notre	notre	PROPN
bracis-33576	337	45	dame	dame	PROPN
bracis-33576	337	46	hesburgh	hesburgh	PROPN
bracis-33576	337	47	library	library	NOUN
bracis-33576	337	48	(	(	PUNCT
bracis-33576	337	49	3000184373	3000184373	NUM
bracis-33576	337	50	)	)	PUNCT
bracis-33576	338	1	©	©	ADP
bracis-33576	338	2	2025	2025	NUM
bracis-33576	338	3	springer	springer	NOUN
bracis-33576	338	4	nature	nature	NOUN