{"id":"https://openalex.org/W7204802369","doi":"https://doi.org/10.18653/2026.findings-emnlp.1300","title":"When Linguistic and Internal Confidence Diverge in Large Language Models","display_name":"When Linguistic and Internal Confidence Diverge in Large Language Models","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7204802369","doi":"https://doi.org/10.18653/2026.findings-emnlp.1300"},"language":"en","primary_location":{"id":"doi:10.18653/2026.findings-emnlp.1300","doi":"https://doi.org/10.18653/2026.findings-emnlp.1300","is_oa":true,"landing_page_url":"https://doi.org/10.18653/2026.findings-emnlp.1300","pdf_url":"https://aclanthology.org/2026.findings-emnlp.1300.pdf","source":{"id":"https://openalex.org/S7407096085","display_name":"Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2026","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","is_traditional":true,"indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"bronze","oa_url":"https://aclanthology.org/2026.findings-emnlp.1300.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5108992299","display_name":"Hefan Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I107672454","display_name":"Dartmouth College","ror":"https://ror.org/049s0rh22","country_code":"US","type":"education","lineage":["https://openalex.org/I107672454"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Hefan Zhang","raw_affiliation_strings":["Dartmouth College ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Dartmouth College ,","institution_ids":["https://openalex.org/I107672454"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125107923","display_name":"Bingquan Zhang","orcid":"https://orcid.org/0000-0003-0182-6595","observed_orcids":["https://orcid.org/0000-0003-0182-6595"]},"institutions":[{"id":"https://openalex.org/I107672454","display_name":"Dartmouth College","ror":"https://ror.org/049s0rh22","country_code":"US","type":"education","lineage":["https://openalex.org/I107672454"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Bingquan Zhang","raw_affiliation_strings":["Dartmouth College ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Dartmouth College ,","institution_ids":["https://openalex.org/I107672454"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148804066","display_name":"Ming Cheng","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I107672454","display_name":"Dartmouth College","ror":"https://ror.org/049s0rh22","country_code":"US","type":"education","lineage":["https://openalex.org/I107672454"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ming Cheng","raw_affiliation_strings":["Dartmouth College ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Dartmouth College ,","institution_ids":["https://openalex.org/I107672454"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5051476268","display_name":"Saeed Hassanpour","orcid":"https://orcid.org/0000-0001-9460-6414","observed_orcids":["https://orcid.org/0000-0001-9460-6414"]},"institutions":[{"id":"https://openalex.org/I107672454","display_name":"Dartmouth College","ror":"https://ror.org/049s0rh22","country_code":"US","type":"education","lineage":["https://openalex.org/I107672454"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Saeed Hassanpour","raw_affiliation_strings":["Dartmouth College ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Dartmouth College ,","institution_ids":["https://openalex.org/I107672454"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5043402333","display_name":"Weicheng Ma","orcid":"https://orcid.org/0000-0001-7494-9874","observed_orcids":["https://orcid.org/0000-0001-7494-9874"]},"institutions":[{"id":"https://openalex.org/I177721651","display_name":"Oakland University","ror":"https://ror.org/01ythxj32","country_code":"US","type":"education","lineage":["https://openalex.org/I177721651"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Weicheng Ma","raw_affiliation_strings":["Oakland University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Oakland University","institution_ids":["https://openalex.org/I177721651"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5035399743","display_name":"Soroush Vosoughi","orcid":"https://orcid.org/0000-0002-2564-8909","observed_orcids":["https://orcid.org/0000-0002-2564-8909"]},"institutions":[{"id":"https://openalex.org/I107672454","display_name":"Dartmouth College","ror":"https://ror.org/049s0rh22","country_code":"US","type":"education","lineage":["https://openalex.org/I107672454"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Soroush Vosoughi","raw_affiliation_strings":["Dartmouth College ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Dartmouth College ,","institution_ids":["https://openalex.org/I107672454"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":2.6448,"has_fulltext":true,"cited_by_count":2,"citation_normalized_percentile":{"value":0.90107229,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":95,"max":96},"biblio":{"volume":null,"issue":null,"first_page":"27492","last_page":"27517"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.42899999022483826,"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":0.42899999022483826,"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":0.2759000062942505,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.1687999963760376,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.8320000171661377},{"id":"https://openalex.org/keywords/linguistic-confidence","display_name":"linguistic confidence","score":0.824999988079071},{"id":"https://openalex.org/keywords/confidence-calibration","display_name":"confidence calibration","score":0.6899999976158142},{"id":"https://openalex.org/keywords/semantic-entropy","display_name":"semantic entropy","score":0.6779999732971191},{"id":"https://openalex.org/keywords/uncertainty-estimation","display_name":"uncertainty estimation","score":0.6129999756813049},{"id":"https://openalex.org/keywords/instruction-tuned-models","display_name":"instruction-tuned models","score":0.7179999947547913},{"id":"https://openalex.org/keywords/prompt-engineering","display_name":"prompt engineering","score":0.9100000262260437},{"id":"https://openalex.org/keywords/confidence-distributions","display_name":"confidence distributions","score":0.6309999823570251}],"concepts":[{"id":"https://openalex.org/C44249647","wikidata":"https://www.wikidata.org/wiki/Q208498","display_name":"Confidence interval","level":2,"score":0.5450999736785889},{"id":"https://openalex.org/C43214815","wikidata":"https://www.wikidata.org/wiki/Q7310987","display_name":"Reliability (semiconductor)","level":3,"score":0.45010000467300415},{"id":"https://openalex.org/C2777267654","wikidata":"https://www.wikidata.org/wiki/Q3519023","display_name":"Test (biology)","level":2,"score":0.41019999980926514},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.39469999074935913},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.38769999146461487},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.3840999901294708},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.3804999887943268},{"id":"https://openalex.org/C13662513","wikidata":"https://www.wikidata.org/wiki/Q5160087","display_name":"Confidence distribution","level":3,"score":0.37299999594688416},{"id":"https://openalex.org/C93518851","wikidata":"https://www.wikidata.org/wiki/Q180160","display_name":"Metadata","level":2,"score":0.36800000071525574},{"id":"https://openalex.org/C2909755999","wikidata":"https://www.wikidata.org/wiki/Q4751126","display_name":"Low Confidence","level":2,"score":0.3643999993801117},{"id":"https://openalex.org/C164226766","wikidata":"https://www.wikidata.org/wiki/Q7293202","display_name":"Rank (graph theory)","level":2,"score":0.3605000078678131},{"id":"https://openalex.org/C2778589607","wikidata":"https://www.wikidata.org/wiki/Q1250335","display_name":"Animacy","level":2,"score":0.3449000120162964},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.32710000872612},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3082999885082245},{"id":"https://openalex.org/C83546350","wikidata":"https://www.wikidata.org/wiki/Q1139051","display_name":"Regression","level":2,"score":0.30550000071525574},{"id":"https://openalex.org/C90329073","wikidata":"https://www.wikidata.org/wiki/Q914232","display_name":"Ask price","level":2,"score":0.30550000071525574},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3021000027656555},{"id":"https://openalex.org/C152877465","wikidata":"https://www.wikidata.org/wiki/Q208042","display_name":"Regression analysis","level":2,"score":0.27639999985694885},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.25440001487731934}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.18653/2026.findings-emnlp.1300","doi":"https://doi.org/10.18653/2026.findings-emnlp.1300","is_oa":true,"landing_page_url":"https://doi.org/10.18653/2026.findings-emnlp.1300","pdf_url":"https://aclanthology.org/2026.findings-emnlp.1300.pdf","source":{"id":"https://openalex.org/S7407096085","display_name":"Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2026","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2608.28382","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.28382","pdf_url":"https://arxiv.org/pdf/2608.28382","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2608.28382","doi":"https://doi.org/10.48550/arxiv.2608.28382","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.28382","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.18653/2026.findings-emnlp.1300","doi":"https://doi.org/10.18653/2026.findings-emnlp.1300","is_oa":true,"landing_page_url":"https://doi.org/10.18653/2026.findings-emnlp.1300","pdf_url":"https://aclanthology.org/2026.findings-emnlp.1300.pdf","source":{"id":"https://openalex.org/S7407096085","display_name":"Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G8672557266","display_name":"Collaborative Research: EAGER: NAIRR Pilot: An AI infrastructure to advance research in reducing poverty and informing social policy","funder_award_id":"2452367","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7204802369.pdf","grobid_xml":"https://content.openalex.org/works/W7204802369.grobid-xml"},"referenced_works_count":27,"referenced_works":["https://openalex.org/W2612431505","https://openalex.org/W2799054028","https://openalex.org/W3035441651","https://openalex.org/W3083410900","https://openalex.org/W3210904070","https://openalex.org/W4281748205","https://openalex.org/W4306808908","https://openalex.org/W4321594373","https://openalex.org/W4322718191","https://openalex.org/W4378945636","https://openalex.org/W4384918448","https://openalex.org/W4387561528","https://openalex.org/W4389520749","https://openalex.org/W4389524305","https://openalex.org/W4390723197","https://openalex.org/W4401042415","https://openalex.org/W4401042444","https://openalex.org/W4401306886","https://openalex.org/W4402671653","https://openalex.org/W4402733601","https://openalex.org/W4404088575","https://openalex.org/W4405655184","https://openalex.org/W4405968021","https://openalex.org/W4412875472","https://openalex.org/W4414939355","https://openalex.org/W4416050792","https://openalex.org/W7055216282"],"related_works":["https://openalex.org/W7083796312","https://openalex.org/W4402667062","https://openalex.org/W4405431306","https://openalex.org/W7170385212","https://openalex.org/W7213589000","https://openalex.org/W4417447091","https://openalex.org/W4399510385","https://openalex.org/W7141135669","https://openalex.org/W4385734246","https://openalex.org/W4404782672"],"abstract_inverted_index":{"Users":[0],"often":[1,87],"ask":[2],"large":[3],"language":[4],"models":[5,38,86],"(LLMs)":[6],"to":[7],"report":[8,88],"how":[9],"confident":[10],"they":[11,97,128],"are,":[12],"but":[13,96,175],"it":[14,76,176],"is":[15,71],"unclear":[16],"whether":[17,61],"such":[18],"linguistic":[19,45,62,163],"confidence":[20,46,49,63,90,101,116,131,139,168,183],"tracks":[21,64],"the":[22,109,144,180],"model's":[23],"internal":[24],"confidence.We":[25],"study":[26],"this":[27],"question":[28],"across":[29],"8":[30],"classification":[31],"tasks,":[32],"2":[33],"generation":[34],"tasks":[35],"and":[36,56,81,91,103],"30":[37],"from":[39],"three":[40,51],"families.For":[41],"classification,":[42],"we":[43,59],"compare":[44],"with":[47,148,188],"logits-based":[48],"along":[50],"axes:":[52],"association,":[53,95],"magnitude":[54],"agreement":[55],"calibration.For":[57],"generation,":[58],"test":[60],"semantic-entropybased":[65],"uncertainty.The":[66],"axes":[67],"frequently":[68],"diverge.Instance-level":[69],"association":[70],"weak":[72],"on":[73,78],"average,":[74],"although":[75],"improves":[77],"easier":[79],"items":[80],"for":[82],"stronger":[83],"base":[84],"models.Instructiontuned":[85],"higher":[89,94],"sometimes":[92],"show":[93,134],"also":[98],"have":[99],"larger":[100],"gaps":[102],"worse":[104],"calibration.Prompt":[105],"design":[106],"mostly":[107],"changes":[108],"distribution":[110,169],"of":[111,138,143,162],"reported":[112],"confidence.Attitude":[113],"cues":[114],"inflate":[115],"without":[117],"improving":[118],"alignment,":[119],"while":[120],"score":[121],"exemplars":[122],"can":[123,170],"preserve":[124],"rank-order":[125],"signal":[126],"when":[127],"avoid":[129],"collapsed":[130],"values.Regression":[132],"analyses":[133],"that":[135],"distributional":[136],"properties":[137],"scores":[140,181],"explain":[141],"much":[142],"observed":[145],"alignment":[146],"pattern,":[147],"model":[149],"metadata":[150],"playing":[151],"a":[152,159],"smaller":[153],"role":[154],"after":[155],"controls.These":[156],"results":[157],"support":[158],"lossy-channel":[160],"view":[161],"confidence.A":[164],"more":[165],"dispersed":[166],"verbal":[167],"carry":[171],"useful":[172],"rank":[173],"information,":[174],"does":[177],"not":[178],"make":[179],"calibrated.Linguistic":[182],"should":[184],"therefore":[185],"be":[186],"evaluated":[187],"multi-axis":[189],"diagnostics":[190],"before":[191],"being":[192],"used":[193],"in":[194],"downstream":[195],"reliability":[196],"pipelines.":[197]},"counts_by_year":[{"year":2026,"cited_by_count":2}],"updated_date":"2026-11-08T12:12:28.789561","created_date":"2026-09-01T00:00:00"}