{"id":"https://openalex.org/W7160934233","doi":"https://doi.org/10.48550/arxiv.2605.08432","title":"A Semantic-Sampling Framework for Evaluating Calibration in Open-Ended Question Answering","display_name":"A Semantic-Sampling Framework for Evaluating Calibration in Open-Ended Question Answering","publication_year":2026,"publication_date":"2026-05-08","ids":{"openalex":"https://openalex.org/W7160934233","doi":"https://doi.org/10.48550/arxiv.2605.08432"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.08432","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.08432","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.2605.08432","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5135928420","display_name":"Zhanliang Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Zhanliang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101368183","display_name":"Jiancong Xiao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiao, Jiancong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135932154","display_name":"Ruochen Jin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jin, Ruochen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135923523","display_name":"Shu Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Shu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5108939216","display_name":"Bojian Hou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hou, Bojian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5135979288","display_name":"Li Shen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shen, Li","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":0.7692000269889832,"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":0.7692000269889832,"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/T13702","display_name":"Machine Learning in Healthcare","score":0.061900001019239426,"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":0.03500000014901161,"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/question-answering","display_name":"Question answering","score":0.7892000079154968},{"id":"https://openalex.org/keywords/calibration","display_name":"Calibration","score":0.7196000218391418},{"id":"https://openalex.org/keywords/estimator","display_name":"Estimator","score":0.6001999974250793},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.5307999849319458},{"id":"https://openalex.org/keywords/software-deployment","display_name":"Software deployment","score":0.5297999978065491},{"id":"https://openalex.org/keywords/confidence-interval","display_name":"Confidence interval","score":0.430400013923645},{"id":"https://openalex.org/keywords/baseline","display_name":"Baseline (sea)","score":0.3986000120639801},{"id":"https://openalex.org/keywords/construct","display_name":"Construct (python library)","score":0.3571000099182129}],"concepts":[{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.7892000079154968},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7519999742507935},{"id":"https://openalex.org/C165838908","wikidata":"https://www.wikidata.org/wiki/Q736777","display_name":"Calibration","level":2,"score":0.7196000218391418},{"id":"https://openalex.org/C185429906","wikidata":"https://www.wikidata.org/wiki/Q1130160","display_name":"Estimator","level":2,"score":0.6001999974250793},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.5307999849319458},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.5297999978065491},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5074999928474426},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.43880000710487366},{"id":"https://openalex.org/C44249647","wikidata":"https://www.wikidata.org/wiki/Q208498","display_name":"Confidence interval","level":2,"score":0.430400013923645},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.40630000829696655},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.40299999713897705},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.3986000120639801},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3628000020980835},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.3571000099182129},{"id":"https://openalex.org/C151201525","wikidata":"https://www.wikidata.org/wiki/Q177239","display_name":"Limit (mathematics)","level":2,"score":0.3479999899864197},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.3224000036716461},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.29350000619888306},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.2870999872684479},{"id":"https://openalex.org/C120936955","wikidata":"https://www.wikidata.org/wiki/Q2155640","display_name":"Empirical research","level":2,"score":0.2745000123977661},{"id":"https://openalex.org/C2778067643","wikidata":"https://www.wikidata.org/wiki/Q166507","display_name":"Interval (graph theory)","level":2,"score":0.27079999446868896},{"id":"https://openalex.org/C166052673","wikidata":"https://www.wikidata.org/wiki/Q83021","display_name":"Empirical evidence","level":2,"score":0.25760000944137573},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.2572999894618988},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.25690001249313354},{"id":"https://openalex.org/C155108698","wikidata":"https://www.wikidata.org/wiki/Q1231081","display_name":"Randomized experiment","level":2,"score":0.25040000677108765}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.08432","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.08432","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.2605.08432","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.08432","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":[{"display_name":"Peace, Justice and strong institutions","score":0.7647984027862549,"id":"https://metadata.un.org/sdg/16"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Calibration":[0,108],"measures":[1],"whether":[2],"a":[3,99,110,149,192],"model's":[4],"predicted":[5],"confidence":[6,84,157,218],"aligns":[7],"with":[8,179],"its":[9],"empirical":[10],"accuracy,":[11],"and":[12,30,80,87,90,128,147,165,212,219],"is":[13,67,85],"central":[14],"to":[15,47],"the":[16,41,59,121,130,143],"reliable":[17],"deployment":[18,62],"of":[19,45],"large":[20],"language":[21],"models":[22],"(LLMs)":[23],"in":[24,50],"high-stakes":[25],"domains":[26],"such":[27],"as":[28,133,191],"medicine":[29],"law.":[31],"While":[32],"much":[33],"recent":[34],"work":[35],"focuses":[36],"on":[37,94,171,176,198],"improving":[38],"LLM":[39],"calibration,":[40],"equally":[42],"important":[43],"question":[44,56,195],"how":[46],"evaluate":[48],"it":[49],"realistic":[51],"settings":[52],"remains":[53],"underdeveloped.":[54],"Open-ended":[55],"answering":[57],"(QA),":[58],"most":[60],"common":[61],"setting":[63],"for":[64,114,194],"modern":[65],"LLMs,":[66],"where":[68],"existing":[69,220],"evaluation":[70,112,226],"methods":[71,92],"fall":[72],"short:":[73],"logit-based":[74,225],"metrics":[75],"need":[76],"restricted":[77],"output":[78],"formats":[79],"internal":[81,228],"probabilities;":[82],"verbalized":[83,217],"self-reported":[86],"often":[88],"overconfident;":[89],"sampling-based":[91,221],"rely":[93],"task-specific":[95],"extraction":[96],"rules":[97],"without":[98],"clear":[100],"finite-sample":[101],"target.":[102],"We":[103,135,159],"introduce":[104],"Sem-ECE":[105,215],"(Semantic-Sampling":[106],"Expected":[107],"Error),":[109],"calibration":[111,184],"framework":[113],"open-ended":[115,200],"QA":[116,201],"that":[117,152,168,214],"samples":[118],"answers":[119],"from":[120,156],"model,":[122],"groups":[123],"them":[124],"into":[125],"semantic":[126],"classes,":[127],"uses":[129],"resulting":[131],"frequencies":[132],"confidence.":[134],"study":[136],"two":[137],"estimators":[138],"within":[139],"this":[140],"framework:":[141],"Sem$_1$-ECE,":[142],"same-sample":[144],"self-consistency":[145],"score,":[146],"Sem$_2$-ECE,":[148],"held-out":[150],"variant":[151],"separates":[153],"answer":[154],"selection":[155],"evaluation.":[158],"prove":[160],"both":[161],"are":[162,230],"asymptotically":[163],"unbiased,":[164],"further":[166],"show":[167,213],"they":[169],"agree":[170],"easy":[172],"questions":[173],"but":[174],"diverge":[175],"hard":[177],"ones":[178],"Sem$_2$":[180],"achieving":[181],"strictly":[182],"smaller":[183],"error,":[185],"so":[186],"their":[187],"gap":[188],"also":[189],"serves":[190],"diagnostic":[193],"difficulty.":[196],"Experiments":[197],"three":[199],"benchmarks":[202],"across":[203],"five":[204],"leading":[205],"commercial":[206],"LLMs":[207],"match":[208],"our":[209],"theoretical":[210],"predictions":[211],"outperforms":[216],"methods,":[222],"while":[223],"complementing":[224],"when":[227],"probabilities":[229],"unavailable.":[231]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-13T00:00:00"}