{"id":"https://openalex.org/W7124142445","doi":"https://doi.org/10.48550/arxiv.2601.07965","title":"When Models Know When They Do Not Know: Calibration, Cascading, and Cleaning","display_name":"When Models Know When They Do Not Know: Calibration, Cascading, and Cleaning","publication_year":2026,"publication_date":"2026-01-12","ids":{"openalex":"https://openalex.org/W7124142445","doi":"https://doi.org/10.48550/arxiv.2601.07965"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2601.07965","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2601.07965","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":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2601.07965","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5123018443","display_name":"Chenjie Hao","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Hao, Chenjie","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123019438","display_name":"Weyl Lu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu, Weyl","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5087580139","display_name":"Yuko Ishiwaka","orcid":"https://orcid.org/0000-0003-2243-3643"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ishiwaka, Yuko","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5045507250","display_name":"Zengyi Li","orcid":"https://orcid.org/0000-0002-5577-3321"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Zengyi","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5085579143","display_name":"Weier Wan","orcid":"https://orcid.org/0000-0003-3507-9014"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wan, Weier","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5123015049","display_name":"Yubei Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Yubei","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5123018443"],"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.24809999763965607,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.24809999763965607,"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":0.19359999895095825,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.1348000019788742,"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/comparability","display_name":"Comparability","score":0.833299994468689},{"id":"https://openalex.org/keywords/calibration","display_name":"Calibration","score":0.6266000270843506},{"id":"https://openalex.org/keywords/exploit","display_name":"Exploit","score":0.6161999702453613},{"id":"https://openalex.org/keywords/reliability","display_name":"Reliability (semiconductor)","score":0.5871999859809875},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.5249000191688538},{"id":"https://openalex.org/keywords/cascade","display_name":"Cascade","score":0.4611000120639801},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.4560999870300293},{"id":"https://openalex.org/keywords/scale","display_name":"Scale (ratio)","score":0.4526999890804291},{"id":"https://openalex.org/keywords/range","display_name":"Range (aeronautics)","score":0.39480000734329224}],"concepts":[{"id":"https://openalex.org/C197947376","wikidata":"https://www.wikidata.org/wiki/Q5155608","display_name":"Comparability","level":2,"score":0.833299994468689},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7588000297546387},{"id":"https://openalex.org/C165838908","wikidata":"https://www.wikidata.org/wiki/Q736777","display_name":"Calibration","level":2,"score":0.6266000270843506},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.6161999702453613},{"id":"https://openalex.org/C43214815","wikidata":"https://www.wikidata.org/wiki/Q7310987","display_name":"Reliability (semiconductor)","level":3,"score":0.5871999859809875},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.526199996471405},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.5249000191688538},{"id":"https://openalex.org/C34146451","wikidata":"https://www.wikidata.org/wiki/Q5048094","display_name":"Cascade","level":2,"score":0.4611000120639801},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4577000141143799},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.4560999870300293},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.4526999890804291},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.41620001196861267},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.39480000734329224},{"id":"https://openalex.org/C74172769","wikidata":"https://www.wikidata.org/wiki/Q1446839","display_name":"Routing (electronic design automation)","level":2,"score":0.367900013923645},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.33649998903274536},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.32280001044273376},{"id":"https://openalex.org/C58489278","wikidata":"https://www.wikidata.org/wiki/Q1172284","display_name":"Data set","level":2,"score":0.32120001316070557},{"id":"https://openalex.org/C16910744","wikidata":"https://www.wikidata.org/wiki/Q7705759","display_name":"Test data","level":2,"score":0.310699999332428},{"id":"https://openalex.org/C200601418","wikidata":"https://www.wikidata.org/wiki/Q2193887","display_name":"Reliability engineering","level":1,"score":0.3041999936103821},{"id":"https://openalex.org/C120936955","wikidata":"https://www.wikidata.org/wiki/Q2155640","display_name":"Empirical research","level":2,"score":0.30149999260902405},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.2935999929904938},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.2809999883174896},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.27070000767707825},{"id":"https://openalex.org/C160920958","wikidata":"https://www.wikidata.org/wiki/Q7662746","display_name":"Synthetic data","level":2,"score":0.26809999346733093},{"id":"https://openalex.org/C133462117","wikidata":"https://www.wikidata.org/wiki/Q4929239","display_name":"Data collection","level":2,"score":0.2671999931335449},{"id":"https://openalex.org/C2777267654","wikidata":"https://www.wikidata.org/wiki/Q3519023","display_name":"Test (biology)","level":2,"score":0.2644999921321869},{"id":"https://openalex.org/C21200559","wikidata":"https://www.wikidata.org/wiki/Q7451068","display_name":"Sensitivity (control systems)","level":2,"score":0.2619999945163727},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.25940001010894775},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.2540999948978424},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.25060001015663147}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2601.07965","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2601.07965","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":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2601.07965","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2601.07965","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":"article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"When":[0],"a":[1,19,64,89,113,126,214,253],"model":[2,20,80,149,161,203],"knows":[3],"when":[4,94,247],"it":[5,24,95],"does":[6,25,96],"not":[7,26,97,250],"know,":[8],"many":[9],"possibilities":[10],"emerge.":[11],"The":[12],"first":[13,100],"question":[14],"is":[15,31,252],"how":[16],"to":[17,21,32,41,73,86,92,109,180,198,226,245],"enable":[18],"recognize":[22,93,246],"that":[23,52,71,220,242],"know.":[27,98],"A":[28],"promising":[29],"approach":[30],"use":[33],"confidence,":[34],"computed":[35],"from":[36,168],"the":[37,120,134,164,169],"model's":[38,90],"internal":[39],"signals,":[40],"reflect":[42],"its":[43],"ignorance.":[44],"Prior":[45],"work":[46],"in":[47,187,230],"specific":[48],"domains":[49],"has":[50],"shown":[51],"calibration":[53],"can":[54],"provide":[55],"reliable":[56],"confidence":[57,107],"estimates.":[58],"In":[59],"this":[60],"work,":[61],"we":[62,144,174,190,212],"propose":[63],"simple,":[65],"effective,":[66],"and":[67,76,83,116,136,155,177,189,208,223,232,260],"universal":[68],"training-free":[69],"method":[70,219],"applies":[72],"both":[74],"vision":[75],"language":[77],"models,":[78],"performing":[79],"calibration,":[81],"cascading,":[82],"data":[84,157],"cleaning":[85,158],"better":[87],"exploit":[88],"ability":[91],"We":[99],"highlight":[101],"two":[102,146,193],"key":[103],"empirical":[104],"observations:":[105],"higher":[106,110],"corresponds":[108],"accuracy":[111],"within":[112],"single":[114],"model,":[115],"models":[117,179,194,244],"calibrated":[118,124,139,152,172,210],"on":[119,125,142,160],"validation":[121],"set":[122],"remain":[123],"held-out":[127],"test":[128],"set.":[129],"These":[130],"findings":[131],"empirically":[132],"establish":[133],"reliability":[135],"comparability":[137,170],"of":[138,171,195],"confidence.":[140],"Building":[141],"this,":[143],"introduce":[145],"applications:":[147],"(1)":[148],"cascading":[150],"with":[151,183],"advantage":[153],"routing":[154,165],"(2)":[156],"based":[159],"ensemble.":[162],"Using":[163],"signal":[166],"derived":[167],"confidences,":[173,211],"cascade":[175,192],"large":[176],"small":[178],"improve":[181],"efficiency":[182],"almost":[184],"no":[185],"compromise":[186],"accuracy,":[188],"further":[191],"comparable":[196],"scale":[197],"achieve":[199],"performance":[200],"beyond":[201],"either":[202],"alone.":[204],"Leveraging":[205],"multiple":[206],"experts":[207],"their":[209],"design":[213],"simple":[215],"yet":[216],"effective":[217],"data-cleaning":[218],"balances":[221],"precision":[222],"detection":[224],"rate":[225],"identify":[227],"mislabeled":[228],"samples":[229],"ImageNet":[231],"Massive":[233],"Multitask":[234],"Language":[235],"Understanding":[236],"(MMLU)":[237],"datasets.":[238],"Our":[239],"results":[240],"demonstrate":[241],"enabling":[243],"they":[248],"do":[249],"know":[251],"practical":[254],"step":[255],"toward":[256],"more":[257],"efficient,":[258],"reliable,":[259],"trustworthy":[261],"AI.":[262]},"counts_by_year":[],"updated_date":"2026-01-15T23:21:31.212559","created_date":"2026-01-15T00:00:00"}