{"id":"https://openalex.org/W7161010363","doi":"https://doi.org/10.48550/arxiv.2605.12446","title":"ORCE: Order-Aware Alignment of Verbalized Confidence in Large Language Models","display_name":"ORCE: Order-Aware Alignment of Verbalized Confidence in Large Language Models","publication_year":2026,"publication_date":"2026-05-12","ids":{"openalex":"https://openalex.org/W7161010363","doi":"https://doi.org/10.48550/arxiv.2605.12446"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.12446","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.12446","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.12446","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5136016662","display_name":"chen li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Chen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136050368","display_name":"Xiaoling Hu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hu, Xiaoling","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5073393494","display_name":"Songzhu Zheng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zheng, Songzhu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136003109","display_name":"Jiawei Zhou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Jiawei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5136052014","display_name":"Chao Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Chao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.7526999711990356,"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":0.7526999711990356,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.04259999841451645,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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":0.026499999687075615,"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/correctness","display_name":"Correctness","score":0.8185999989509583},{"id":"https://openalex.org/keywords/low-confidence","display_name":"Low Confidence","score":0.5027999877929688},{"id":"https://openalex.org/keywords/certainty","display_name":"Certainty","score":0.48339998722076416},{"id":"https://openalex.org/keywords/decoupling","display_name":"Decoupling (probability)","score":0.47380000352859497},{"id":"https://openalex.org/keywords/confidence-interval","display_name":"Confidence interval","score":0.4408000111579895},{"id":"https://openalex.org/keywords/construct","display_name":"Construct (python library)","score":0.43290001153945923},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.3946000039577484},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.3935999870300293}],"concepts":[{"id":"https://openalex.org/C55439883","wikidata":"https://www.wikidata.org/wiki/Q360812","display_name":"Correctness","level":2,"score":0.8185999989509583},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7185999751091003},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5375999808311462},{"id":"https://openalex.org/C2909755999","wikidata":"https://www.wikidata.org/wiki/Q4751126","display_name":"Low Confidence","level":2,"score":0.5027999877929688},{"id":"https://openalex.org/C7493553","wikidata":"https://www.wikidata.org/wiki/Q1520777","display_name":"Certainty","level":2,"score":0.48339998722076416},{"id":"https://openalex.org/C205606062","wikidata":"https://www.wikidata.org/wiki/Q5249645","display_name":"Decoupling (probability)","level":2,"score":0.47380000352859497},{"id":"https://openalex.org/C44249647","wikidata":"https://www.wikidata.org/wiki/Q208498","display_name":"Confidence interval","level":2,"score":0.4408000111579895},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.43290001153945923},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.42179998755455017},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.3946000039577484},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.3935999870300293},{"id":"https://openalex.org/C165838908","wikidata":"https://www.wikidata.org/wiki/Q736777","display_name":"Calibration","level":2,"score":0.38370001316070557},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3398999869823456},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.322299987077713},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.2870999872684479},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.2825999855995178},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.27869999408721924},{"id":"https://openalex.org/C2776187449","wikidata":"https://www.wikidata.org/wiki/Q1513879","display_name":"Natural language generation","level":3,"score":0.267300009727478},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.2513999938964844},{"id":"https://openalex.org/C41426520","wikidata":"https://www.wikidata.org/wiki/Q1192065","display_name":"Point estimation","level":2,"score":0.25060001015663147}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.12446","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.12446","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.12446","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.12446","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"language":[1],"models":[2,28],"(LLMs)":[3],"often":[4,57],"produce":[5],"answers":[6],"with":[7,72,117,138],"high":[8],"certainty":[9],"even":[10,47],"when":[11,48],"they":[12],"are":[13,51],"incorrect,":[14],"making":[15],"reliable":[16],"confidence":[17,32,62,87,98,106,116,174,182,193],"estimation":[18,183],"essential":[19],"for":[20,85],"deployment":[21],"in":[22,33],"real-world":[23],"scenarios.":[24],"Verbalized":[25],"confidence,":[26],"where":[27],"explicitly":[29],"state":[30],"their":[31],"natural":[34],"language,":[35],"provides":[36],"a":[37,80,122],"flexible":[38],"and":[39,61,82,95,129,151,160,187],"user-facing":[40],"uncertainty":[41],"signal":[42],"that":[43,135,155,172],"can":[44,66,175],"be":[45,176],"applied":[46],"token":[49],"logits":[50],"unavailable.":[52],"However,":[53],"existing":[54],"verbalized-confidence":[55],"methods":[56],"optimize":[58,130],"answer":[59,73,94,167,185],"generation":[60,63,186],"jointly,":[64],"which":[65],"cause":[67],"confidence-alignment":[68],"objectives":[69,134],"to":[70,143],"interfere":[71],"accuracy.":[74,168],"In":[75],"this":[76],"work,":[77],"we":[78,120],"propose":[79],"decoupled":[81],"order-aware":[83],"framework":[84],"verbalized":[86,146,173],"calibration.":[88],"Our":[89],"method":[90,157],"first":[91],"generates":[92],"an":[93],"then":[96],"estimates":[97],"conditioned":[99],"on":[100,149],"the":[101,111,189],"fixed":[102],"question--answer":[103],"pair,":[104],"allowing":[105],"optimization":[107],"without":[108],"directly":[109],"perturbing":[110],"answer-generation":[112],"process.":[113],"To":[114],"align":[115],"correctness":[118,141],"likelihood,":[119],"construct":[121],"sampling-based":[123],"surrogate":[124],"from":[125,184],"multiple":[126],"model":[127],"completions":[128],"rank-based":[131],"reinforcement":[132],"learning":[133],"encourage":[136],"responses":[137],"higher":[139,145],"estimated":[140],"likelihood":[142],"receive":[144],"confidence.":[147],"Experiments":[148],"reasoning":[150],"knowledge-intensive":[152],"benchmarks":[153],"show":[154],"our":[156],"improves":[158],"calibration":[159],"failure":[161],"prediction":[162],"performance":[163],"while":[164],"largely":[165],"preserving":[166],"These":[169],"results":[170],"demonstrate":[171],"more":[177],"reliably":[178],"aligned":[179],"by":[180],"decoupling":[181],"optimizing":[188],"relative":[190],"ordering":[191],"of":[192],"across":[194],"responses.":[195]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-14T00:00:00"}