{"id":"https://openalex.org/W7154453268","doi":"https://doi.org/10.48550/arxiv.2604.12046","title":"Think Through Uncertainty: Improving Long-Form Generation Factuality via Reasoning Calibration","display_name":"Think Through Uncertainty: Improving Long-Form Generation Factuality via Reasoning Calibration","publication_year":2026,"publication_date":"2026-04-13","ids":{"openalex":"https://openalex.org/W7154453268","doi":"https://doi.org/10.48550/arxiv.2604.12046"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.12046","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.12046","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":"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2604.12046","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5133640468","display_name":"Xin Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Xin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5133679514","display_name":"Lu Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Lu","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":0.4763999879360199,"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":0.4763999879360199,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.14480000734329224,"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/T11636","display_name":"Artificial Intelligence in Healthcare and Education","score":0.05849999934434891,"subfield":{"id":"https://openalex.org/subfields/2718","display_name":"Health Informatics"},"field":{"id":"https://openalex.org/fields/27","display_name":"Medicine"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/correctness","display_name":"Correctness","score":0.6880999803543091},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6348000168800354},{"id":"https://openalex.org/keywords/calibration","display_name":"Calibration","score":0.5751000046730042},{"id":"https://openalex.org/keywords/hallucinating","display_name":"Hallucinating","score":0.5152000188827515},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.4950000047683716},{"id":"https://openalex.org/keywords/low-confidence","display_name":"Low Confidence","score":0.4104999899864197}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.73580002784729},{"id":"https://openalex.org/C55439883","wikidata":"https://www.wikidata.org/wiki/Q360812","display_name":"Correctness","level":2,"score":0.6880999803543091},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6348000168800354},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6323000192642212},{"id":"https://openalex.org/C165838908","wikidata":"https://www.wikidata.org/wiki/Q736777","display_name":"Calibration","level":2,"score":0.5751000046730042},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5170000195503235},{"id":"https://openalex.org/C2911011789","wikidata":"https://www.wikidata.org/wiki/Q130741","display_name":"Hallucinating","level":2,"score":0.5152000188827515},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.4950000047683716},{"id":"https://openalex.org/C2909755999","wikidata":"https://www.wikidata.org/wiki/Q4751126","display_name":"Low Confidence","level":2,"score":0.4104999899864197},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.38040000200271606},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3540000021457672},{"id":"https://openalex.org/C195344581","wikidata":"https://www.wikidata.org/wiki/Q2555318","display_name":"Automated reasoning","level":2,"score":0.32330000400543213},{"id":"https://openalex.org/C44249647","wikidata":"https://www.wikidata.org/wiki/Q208498","display_name":"Confidence interval","level":2,"score":0.3131999969482422},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.29600000381469727},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.28769999742507935}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.12046","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.12046","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2604.12046","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.12046","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":"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"},"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":[2,43],"(LLMs)":[3],"often":[4],"hallucinate":[5],"in":[6,50,55,230],"long-form":[7,93,112,184],"generation.":[8,216],"Existing":[9],"approaches":[10,77],"mainly":[11],"improve":[12],"factuality":[13,113,185],"through":[14],"post-hoc":[15],"revision":[16],"or":[17],"reinforcement":[18],"learning":[19],"(RL)":[20],"with":[21,139,154],"correctness-based":[22],"rewards,":[23],"but":[24],"they":[25],"do":[26],"not":[27],"teach":[28],"the":[29,86,122,171],"model":[30,152,172],"to":[31,66,80,117,173,212],"estimate":[32,67],"which":[33,89,132],"parts":[34],"of":[35],"its":[36],"generation":[37,94],"are":[38,219],"reliable.":[39],"As":[40],"a":[41,81,108,128,146,227],"result,":[42],"may":[44],"still":[45],"state":[46],"incorrect":[47],"claims":[48,137,177],"confidently":[49],"their":[51],"responses.":[52],"Recent":[53],"advances":[54],"reasoning":[56],"have":[57,63],"significantly":[58],"improved":[59,222],"LLM":[60],"performance,":[61],"and":[62,157,197],"been":[64],"leveraged":[65],"confidence":[68,84,141,153,165],"by":[69,114,210,221,226],"incorporating":[70],"calibration":[71],"into":[72,135],"RL":[73,198],"objectives.":[74],"However,":[75],"existing":[76],"remain":[78],"limited":[79],"single":[82],"scalar":[83],"for":[85,92],"entire":[87],"response,":[88],"is":[90],"insufficient":[91],"where":[95],"uncertainty":[96,120],"varies":[97],"across":[98],"individual":[99],"claims.":[100],"To":[101],"mitigate":[102],"this":[103],"problem,":[104],"we":[105],"propose":[106],"CURE,":[107],"framework":[109],"that":[110,150,188],"improves":[111,191,207],"teaching":[115],"LLMs":[116],"reason":[118],"about":[119],"at":[121,178],"claim":[123],"level.":[124],"We":[125,143],"first":[126],"introduce":[127],"Claim-Aware":[129],"Reasoning":[130],"Protocol,":[131],"structures":[133],"outputs":[134],"atomic":[136],"paired":[138],"explicit":[140],"estimates.":[142],"then":[144,158],"develop":[145],"multi-stage":[147],"training":[148],"pipeline":[149],"aligns":[151],"claims'":[155],"correctness":[156],"optimizes":[159],"on":[160,182,214,232],"factuality.":[161],"The":[162],"resulting":[163],"calibrated":[164],"further":[166],"enables":[167],"selective":[168],"prediction,":[169],"allowing":[170],"abstain":[174],"from":[175],"uncertain":[176],"inference":[179],"time.":[180],"Experiments":[181],"four":[183],"benchmarks":[186],"show":[187],"CURE":[189],"consistently":[190],"factual":[192,202],"accuracy":[193,209],"over":[194],"competitive":[195],"supervised":[196],"baselines,":[199],"while":[200],"maintaining":[201],"recall.":[203],"In":[204],"particular,":[205],"it":[206],"claim-level":[208],"up":[211],"39.9%":[213],"Biography":[215],"These":[217],"gains":[218],"accompanied":[220],"calibration,":[223],"as":[224],"reflected":[225],"16.0%":[228],"increase":[229],"AUROC":[231],"FactBench.":[233]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-04-16T00:00:00"}