{"id":"https://openalex.org/W7160982125","doi":"https://doi.org/10.48550/arxiv.2605.11845","title":"Probabilistic Calibration Is a Trainable Capability in Language Models","display_name":"Probabilistic Calibration Is a Trainable Capability in Language Models","publication_year":2026,"publication_date":"2026-05-12","ids":{"openalex":"https://openalex.org/W7160982125","doi":"https://doi.org/10.48550/arxiv.2605.11845"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.11845","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.11845","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.2605.11845","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5136049764","display_name":"Davide Baldelli","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Baldelli, Davide","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5018930589","display_name":"Sruthi Kuriakose","orcid":"https://orcid.org/0000-0001-6221-7203"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kuriakose, Sruthi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136083684","display_name":"Maryam Hashemzadeh","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hashemzadeh, Maryam","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5075664559","display_name":"Amal Zouaq","orcid":"https://orcid.org/0000-0002-4791-0752"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zouaq, Amal","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5136067183","display_name":"Sarath Chandar","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chandar, Sarath","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":0.5764999985694885,"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":0.5764999985694885,"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":0.08389999717473984,"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"}},{"id":"https://openalex.org/T13629","display_name":"Text Readability and Simplification","score":0.06629999727010727,"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/probabilistic-logic","display_name":"Probabilistic logic","score":0.7415000200271606},{"id":"https://openalex.org/keywords/calibration","display_name":"Calibration","score":0.729200005531311},{"id":"https://openalex.org/keywords/randomness","display_name":"Randomness","score":0.6848999857902527},{"id":"https://openalex.org/keywords/fidelity","display_name":"Fidelity","score":0.6194000244140625},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.557699978351593},{"id":"https://openalex.org/keywords/probability-distribution","display_name":"Probability distribution","score":0.46380001306533813},{"id":"https://openalex.org/keywords/statistical-model","display_name":"Statistical model","score":0.4300999939441681},{"id":"https://openalex.org/keywords/sampling","display_name":"Sampling (signal processing)","score":0.42179998755455017},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.38440001010894775}],"concepts":[{"id":"https://openalex.org/C49937458","wikidata":"https://www.wikidata.org/wiki/Q2599292","display_name":"Probabilistic logic","level":2,"score":0.7415000200271606},{"id":"https://openalex.org/C165838908","wikidata":"https://www.wikidata.org/wiki/Q736777","display_name":"Calibration","level":2,"score":0.729200005531311},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7085999846458435},{"id":"https://openalex.org/C125112378","wikidata":"https://www.wikidata.org/wiki/Q176640","display_name":"Randomness","level":2,"score":0.6848999857902527},{"id":"https://openalex.org/C2776459999","wikidata":"https://www.wikidata.org/wiki/Q2119376","display_name":"Fidelity","level":2,"score":0.6194000244140625},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.557699978351593},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4968999922275543},{"id":"https://openalex.org/C149441793","wikidata":"https://www.wikidata.org/wiki/Q200726","display_name":"Probability distribution","level":2,"score":0.46380001306533813},{"id":"https://openalex.org/C114289077","wikidata":"https://www.wikidata.org/wiki/Q3284399","display_name":"Statistical model","level":2,"score":0.4300999939441681},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.42179998755455017},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4009999930858612},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.396699994802475},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.38440001010894775},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.35929998755455017},{"id":"https://openalex.org/C113364801","wikidata":"https://www.wikidata.org/wiki/Q26674","display_name":"High fidelity","level":2,"score":0.3434000015258789},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.3319000005722046},{"id":"https://openalex.org/C110121322","wikidata":"https://www.wikidata.org/wiki/Q865811","display_name":"Distribution (mathematics)","level":2,"score":0.3070000112056732},{"id":"https://openalex.org/C52740198","wikidata":"https://www.wikidata.org/wiki/Q1539564","display_name":"Importance sampling","level":3,"score":0.301800012588501},{"id":"https://openalex.org/C127491075","wikidata":"https://www.wikidata.org/wiki/Q7617825","display_name":"Stochastic modelling","level":2,"score":0.3005000054836273},{"id":"https://openalex.org/C107673813","wikidata":"https://www.wikidata.org/wiki/Q812534","display_name":"Bayesian probability","level":2,"score":0.29589998722076416},{"id":"https://openalex.org/C8272713","wikidata":"https://www.wikidata.org/wiki/Q176737","display_name":"Stochastic process","level":2,"score":0.2802000045776367},{"id":"https://openalex.org/C61224824","wikidata":"https://www.wikidata.org/wiki/Q2260434","display_name":"Mixture model","level":2,"score":0.27480000257492065},{"id":"https://openalex.org/C196083921","wikidata":"https://www.wikidata.org/wiki/Q7915758","display_name":"Variance (accounting)","level":2,"score":0.272599995136261},{"id":"https://openalex.org/C160920958","wikidata":"https://www.wikidata.org/wiki/Q7662746","display_name":"Synthetic data","level":2,"score":0.2671999931335449},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.25839999318122864},{"id":"https://openalex.org/C149717495","wikidata":"https://www.wikidata.org/wiki/Q117806","display_name":"Gamma distribution","level":2,"score":0.2565000057220459},{"id":"https://openalex.org/C160234255","wikidata":"https://www.wikidata.org/wiki/Q812535","display_name":"Bayesian inference","level":3,"score":0.2542000114917755}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.11845","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.11845","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.2605.11845","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.11845","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":{"Language":[0],"models":[1,40,85],"are":[2,18],"increasingly":[3],"used":[4],"in":[5],"settings":[6],"where":[7],"outputs":[8],"must":[9],"satisfy":[10],"user-specified":[11],"randomness":[12],"constraints,":[13],"yet":[14],"their":[15],"generation":[16,140],"probabilities":[17],"often":[19,126],"poorly":[20],"calibrated":[21],"to":[22],"those":[23],"targets.":[24],"We":[25],"study":[26],"whether":[27],"this":[28],"capability":[29],"can":[30,172],"be":[31,173],"improved":[32,174],"directly":[33],"through":[34,175],"fine-tuning.":[35],"Concretely,":[36],"we":[37],"fine-tune":[38],"language":[39],"on":[41,75,95,128,137],"synthetic":[42],"prompts":[43],"that":[44,59,73,104,169],"require":[45],"sampling":[46],"from":[47,78],"mathematical":[48],"distributions,":[49],"and":[50,69,99,149,185],"compare":[51],"two":[52,117],"Calibration":[53],"Fine-Tuning":[54],"variants:":[55],"a":[56,70,108],"soft-target":[57,133,187],"method":[58,72],"converts":[60],"the":[61,79,116],"desired":[62],"output":[63],"distribution":[64,97],"into":[65],"trie-derived":[66],"next-token":[67],"targets,":[68],"hard-target":[71,123,179],"trains":[74],"sampled":[76],"completions":[77],"same":[80],"target":[81],"distribution.":[82],"Across":[83],"12":[84],"spanning":[86],"four":[87],"families,":[88],"both":[89],"methods":[90,118],"substantially":[91],"improve":[92],"structured-sampling":[93],"fidelity":[94,184],"held-out":[96],"families":[98],"unseen":[100],"parameter":[101],"settings,":[102],"showing":[103],"probabilistic":[105,170],"calibration":[106,171],"is":[107,125,194],"trainable":[109],"capability.":[110],"Under":[111],"our":[112,166,178,186],"selected":[113],"training":[114],"configurations,":[115],"exhibit":[119],"different":[120],"empirical":[121],"profiles:":[122],"fine-tuning":[124,134],"strongest":[127],"structured":[129],"numeric":[130,183],"sampling,":[131],"while":[132],"performs":[135],"better":[136],"broader":[138,190],"stochastic":[139,191],"benchmarks,":[141],"including":[142],"open-ended":[143],"random":[144],"generation,":[145],"multiple-choice":[146],"answer-position":[147],"balancing,":[148],"NoveltyBench.":[150],"The":[151],"gains":[152],"sometimes":[153],"reduce":[154],"downstream":[155],"capability,":[156],"especially":[157],"arithmetic":[158],"reasoning,":[159],"with":[160,177],"costs":[161],"varying":[162],"by":[163],"model.":[164],"Overall,":[165],"results":[167],"show":[168],"fine-tuning,":[176],"configuration":[180,188],"favoring":[181,189],"exact":[182],"transfer.":[192],"Code":[193],"available":[195],"at":[196],"https://github.com/chandar-lab/calibration-finetuning.":[197]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-14T00:00:00"}