{"id":"https://openalex.org/W4412889555","doi":"https://doi.org/10.18653/v1/2025.acl-short.18","title":"Revisiting Epistemic Markers in Confidence Estimation: Can Markers Accurately Reflect Large Language Models\u2019 Uncertainty?","display_name":"Revisiting Epistemic Markers in Confidence Estimation: Can Markers Accurately Reflect Large Language Models\u2019 Uncertainty?","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412889555","doi":"https://doi.org/10.18653/v1/2025.acl-short.18"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.acl-short.18","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-short.18","pdf_url":"https://aclanthology.org/2025.acl-short.18.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.acl-short.18.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5030410270","display_name":"Jiayu Liu","orcid":"https://orcid.org/0000-0001-8639-3308"},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Jiayu Liu","raw_affiliation_strings":["Department of Computer Science and Engineering , HKUST , Hong Kong SAR , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering , HKUST , Hong Kong SAR , China","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5093047682","display_name":"Qing Zong","orcid":null},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Qing Zong","raw_affiliation_strings":["Department of Computer Science and Engineering , HKUST , Hong Kong SAR , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering , HKUST , Hong Kong SAR , China","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100601723","display_name":"Weiqi Wang","orcid":"https://orcid.org/0000-0001-5195-9391"},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Weiqi Wang","raw_affiliation_strings":["Department of Computer Science and Engineering , HKUST , Hong Kong SAR , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering , HKUST , Hong Kong SAR , China","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5020880385","display_name":"Yangqiu Song","orcid":"https://orcid.org/0000-0002-7818-6090"},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Yangqiu Song","raw_affiliation_strings":["Department of Computer Science and Engineering , HKUST , Hong Kong SAR , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering , HKUST , Hong Kong SAR , China","institution_ids":["https://openalex.org/I200769079"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I200769079"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"206","last_page":"221"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9218999743461609,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9218999743461609,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.9140999913215637,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T13910","display_name":"Computational and Text Analysis Methods","score":0.9075000286102295,"subfield":{"id":"https://openalex.org/subfields/3300","display_name":"General Social Sciences"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.5628918409347534},{"id":"https://openalex.org/keywords/estimation","display_name":"Estimation","score":0.5234653353691101},{"id":"https://openalex.org/keywords/confidence-interval","display_name":"Confidence interval","score":0.4512752592563629},{"id":"https://openalex.org/keywords/uncertainty-quantification","display_name":"Uncertainty quantification","score":0.43628332018852234},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4239654541015625},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.3999063968658447},{"id":"https://openalex.org/keywords/epistemology","display_name":"Epistemology","score":0.3222203850746155},{"id":"https://openalex.org/keywords/statistics","display_name":"Statistics","score":0.3197448253631592},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.2520751953125},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.22389647364616394},{"id":"https://openalex.org/keywords/philosophy","display_name":"Philosophy","score":0.156284898519516},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.1264697015285492}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5628918409347534},{"id":"https://openalex.org/C96250715","wikidata":"https://www.wikidata.org/wiki/Q965330","display_name":"Estimation","level":2,"score":0.5234653353691101},{"id":"https://openalex.org/C44249647","wikidata":"https://www.wikidata.org/wiki/Q208498","display_name":"Confidence interval","level":2,"score":0.4512752592563629},{"id":"https://openalex.org/C32230216","wikidata":"https://www.wikidata.org/wiki/Q7882499","display_name":"Uncertainty quantification","level":2,"score":0.43628332018852234},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4239654541015625},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3999063968658447},{"id":"https://openalex.org/C111472728","wikidata":"https://www.wikidata.org/wiki/Q9471","display_name":"Epistemology","level":1,"score":0.3222203850746155},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.3197448253631592},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2520751953125},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.22389647364616394},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.156284898519516},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.1264697015285492},{"id":"https://openalex.org/C201995342","wikidata":"https://www.wikidata.org/wiki/Q682496","display_name":"Systems engineering","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.18653/v1/2025.acl-short.18","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-short.18","pdf_url":"https://aclanthology.org/2025.acl-short.18.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)","raw_type":"proceedings-article"},{"id":"pmh:oai:repository.hkust.edu.hk:1783.1-168675","is_oa":false,"landing_page_url":"http://repository.hkust.edu.hk/ir/Record/1783.1-168675","pdf_url":null,"source":{"id":"https://openalex.org/S4306401796","display_name":"Rare & Special e-Zone (The Hong Kong University of Science and Technology)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I200769079","host_organization_name":"Hong Kong University of Science and Technology","host_organization_lineage":["https://openalex.org/I200769079"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Conference paper"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.acl-short.18","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-short.18","pdf_url":"https://aclanthology.org/2025.acl-short.18.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[{"score":0.41999998688697815,"display_name":"Reduced inequalities","id":"https://metadata.un.org/sdg/10"}],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412889555.pdf","grobid_xml":"https://content.openalex.org/works/W4412889555.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W2995364151","https://openalex.org/W797163146","https://openalex.org/W2068023914","https://openalex.org/W1899429405","https://openalex.org/W2401903608","https://openalex.org/W4221078923","https://openalex.org/W2066458123","https://openalex.org/W3005777437","https://openalex.org/W1578234703","https://openalex.org/W2084691705"],"abstract_inverted_index":{"As":[0],"Large":[1],"Language":[2],"Models":[3],"(LLMs)":[4],"are":[5],"increasingly":[6],"used":[7],"in":[8,81,107],"high-stakes":[9],"domains,":[10],"accurately":[11],"assessing":[12],"their":[13,41,103],"confidence":[14,19,43,62,104,121,132],"is":[15,105],"crucial.Humans":[16],"typically":[17],"express":[18],"through":[20],"epistemic":[21,72,118],"markers":[22,38,96,119],"(e.g.,":[23],"\"fairly":[24],"confident\")":[25],"instead":[26],"of":[27,48,117],"numerical":[28],"values.However,":[29],"it":[30],"remains":[31],"unclear":[32],"whether":[33],"LLMs":[34],"reliably":[35],"use":[36],"these":[37],"to":[39,45],"reflect":[40],"intrinsic":[42],"due":[44],"the":[46,64,100,115,124],"difficulty":[47],"quantifying":[49],"uncertainty":[50],"associated":[51],"with":[52],"various":[53],"markers.To":[54],"address":[55],"this":[56],"gap,":[57],"we":[58],"first":[59],"define":[60],"marker":[61,130],"as":[63],"observed":[65],"accuracy":[66],"when":[67],"a":[68],"model":[69,135],"employs":[70],"an":[71],"marker.We":[73],"evaluate":[74],"its":[75],"stability":[76],"across":[77],"multiple":[78],"question-answering":[79],"datasets":[80],"both":[82],"indistribution":[83],"and":[84,89,133],"out-of-distribution":[85,108],"settings":[86],"for":[87,120,126],"open-source":[88],"proprietary":[90],"LLMs.Our":[91],"results":[92],"show":[93],"that":[94],"while":[95],"generalize":[97],"well":[98],"within":[99],"same":[101],"distribution,":[102],"inconsistent":[106],"scenarios.These":[109],"findings":[110],"raise":[111],"significant":[112],"concerns":[113],"about":[114],"reliability":[116],"estimation,":[122],"underscoring":[123],"need":[125],"improved":[127],"alignment":[128],"between":[129],"based":[131],"actual":[134],"uncertainty.":[136]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}