{"id":"https://openalex.org/W7204178933","doi":"https://doi.org/10.48550/arxiv.2608.22887","title":"Language Models Give Group-Linked Fields the Wrong Weight","display_name":"Language Models Give Group-Linked Fields the Wrong Weight","publication_year":2026,"publication_date":"2026-08-24","ids":{"openalex":"https://openalex.org/W7204178933","doi":"https://doi.org/10.48550/arxiv.2608.22887"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.22887","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.22887","pdf_url":"https://arxiv.org/pdf/2608.22887","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2608.22887","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5137656146","display_name":"Zengqing Wu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I98285908","display_name":"The University of Osaka","ror":"https://ror.org/035t8zc32","country_code":"JP","type":"education","lineage":["https://openalex.org/I98285908"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Wu, Zengqing","raw_affiliation_strings":["University of Osaka Osaka , Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Osaka Osaka , Japan","institution_ids":["https://openalex.org/I98285908"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5148049800","display_name":"Chuan Xiao","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I98285908","display_name":"The University of Osaka","ror":"https://ror.org/035t8zc32","country_code":"JP","type":"education","lineage":["https://openalex.org/I98285908"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Xiao, Chuan","raw_affiliation_strings":["University of Osaka Osaka , Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Osaka Osaka , Japan","institution_ids":["https://openalex.org/I98285908"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I98285908"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"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/T10883","display_name":"Ethics and Social Impacts of AI","score":0.9560999870300293,"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/T10883","display_name":"Ethics and Social Impacts of AI","score":0.9560999870300293,"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":0.03889999911189079,"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/T13643","display_name":"Artificial Intelligence in Law","score":0.0005000000237487257,"subfield":{"id":"https://openalex.org/subfields/3308","display_name":"Law"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.8429999947547913},{"id":"https://openalex.org/keywords/algorithmic-fairness","display_name":"algorithmic fairness","score":0.6129999756813049},{"id":"https://openalex.org/keywords/protected-characteristics","display_name":"protected characteristics","score":0.7360000014305115},{"id":"https://openalex.org/keywords/bayesian-linear-regression","display_name":"Bayesian linear regression","score":0.5979999899864197},{"id":"https://openalex.org/keywords/ranking-accuracy","display_name":"ranking accuracy","score":0.5460000038146973},{"id":"https://openalex.org/keywords/algorithmic-auditing","display_name":"algorithmic auditing","score":0.5059999823570251}],"concepts":[{"id":"https://openalex.org/C2780148112","wikidata":"https://www.wikidata.org/wiki/Q1432581","display_name":"Proxy (statistics)","level":2,"score":0.836899995803833},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5149999856948853},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.42989999055862427},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.42739999294281006},{"id":"https://openalex.org/C162118730","wikidata":"https://www.wikidata.org/wiki/Q1128453","display_name":"Actuarial science","level":1,"score":0.42239999771118164},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.3995000123977661},{"id":"https://openalex.org/C158600405","wikidata":"https://www.wikidata.org/wiki/Q5054566","display_name":"Causal inference","level":2,"score":0.39239999651908875},{"id":"https://openalex.org/C199521495","wikidata":"https://www.wikidata.org/wiki/Q181487","display_name":"Audit","level":2,"score":0.3763999938964844},{"id":"https://openalex.org/C2777120189","wikidata":"https://www.wikidata.org/wiki/Q780067","display_name":"Triage","level":2,"score":0.34689998626708984},{"id":"https://openalex.org/C2780084366","wikidata":"https://www.wikidata.org/wiki/Q37732","display_name":"Demographics","level":2,"score":0.34310001134872437},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.3418000042438507},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.3377000093460083},{"id":"https://openalex.org/C90329073","wikidata":"https://www.wikidata.org/wiki/Q914232","display_name":"Ask price","level":2,"score":0.3269999921321869},{"id":"https://openalex.org/C9652623","wikidata":"https://www.wikidata.org/wiki/Q190109","display_name":"Field (mathematics)","level":2,"score":0.3167000114917755},{"id":"https://openalex.org/C2779458634","wikidata":"https://www.wikidata.org/wiki/Q24963715","display_name":"Debiasing","level":2,"score":0.30379998683929443},{"id":"https://openalex.org/C84525736","wikidata":"https://www.wikidata.org/wiki/Q831366","display_name":"Decision tree","level":2,"score":0.26089999079704285},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.258899986743927},{"id":"https://openalex.org/C11671645","wikidata":"https://www.wikidata.org/wiki/Q5054567","display_name":"Causal model","level":2,"score":0.2583000063896179},{"id":"https://openalex.org/C77805123","wikidata":"https://www.wikidata.org/wiki/Q161272","display_name":"Social psychology","level":1,"score":0.2574000060558319}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.22887","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.22887","pdf_url":"https://arxiv.org/pdf/2608.22887","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2608.22887","doi":"https://doi.org/10.48550/arxiv.2608.22887","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.22887","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2608.22887","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.22887","pdf_url":"https://arxiv.org/pdf/2608.22887","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[{"display_name":"Reduced inequalities","score":0.4000000059604645,"id":"https://metadata.un.org/sdg/10"}],"study_designs":[],"awards":[{"id":"https://openalex.org/G2489078468","display_name":"Research on Enhancing the Performance of Deep Learning and Search for Large-Scale Graphs","funder_award_id":"25H01117","funder_id":"https://openalex.org/F4320334764","funder_display_name":"Japan Society for the Promotion of Science"},{"id":"https://openalex.org/G2781111794","display_name":"\u81ea\u5f8b\u5354\u8abf\u578b\u30c7\u30fc\u30bf\u30d9\u30fc\u30b9\u7d71\u5408\u57fa\u76e4\u306b\u95a2\u3059\u308b\u7814\u7a76","funder_award_id":"23K17456","funder_id":"https://openalex.org/F4320334764","funder_display_name":"Japan Society for the Promotion of Science"},{"id":"https://openalex.org/G4215919285","display_name":"Legal AI for Democratizing the Analysis of Judicial Decisions","funder_award_id":"26K03246","funder_id":"https://openalex.org/F4320334764","funder_display_name":"Japan Society for the Promotion of Science"},{"id":"https://openalex.org/G451006945","display_name":null,"funder_award_id":"JPMJCR22M2","funder_id":"https://openalex.org/F4320334764","funder_display_name":"Japan Society for the Promotion of Science"},{"id":"https://openalex.org/G5026330598","display_name":null,"funder_award_id":"JP-MJCR22M2","funder_id":"https://openalex.org/F4320338075","funder_display_name":"Core Research for Evolutional Science and Technology"},{"id":"https://openalex.org/G5617894317","display_name":"Self-optimized graph database management system with deep learning","funder_award_id":"23K28096","funder_id":"https://openalex.org/F4320334764","funder_display_name":"Japan Society for the Promotion of Science"}],"funders":[{"id":"https://openalex.org/F4320334764","display_name":"Japan Society for the Promotion of Science","ror":"https://ror.org/00hhkn466"},{"id":"https://openalex.org/F4320338075","display_name":"Core Research for Evolutional Science and Technology","ror":"https://ror.org/00097mb19"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7204178933.pdf","grobid_xml":"https://content.openalex.org/works/W7204178933.grobid-xml"},"referenced_works_count":12,"referenced_works":["https://openalex.org/W4311991106","https://openalex.org/W4361020574","https://openalex.org/W4389471473","https://openalex.org/W4391389316","https://openalex.org/W4399317891","https://openalex.org/W4417239928","https://openalex.org/W7125487705","https://openalex.org/W7140331071","https://openalex.org/W7160427237","https://openalex.org/W7160848286","https://openalex.org/W7166717775","https://openalex.org/W7166797992"],"related_works":["https://openalex.org/W7212033163","https://openalex.org/W7223346590","https://openalex.org/W7216127427","https://openalex.org/W7212351249","https://openalex.org/W7223379858","https://openalex.org/W4408342414","https://openalex.org/W7214125621","https://openalex.org/W7223018854","https://openalex.org/W7171809937","https://openalex.org/W7213398409"],"abstract_inverted_index":{"Large":[0],"language":[1],"models":[2,89,109],"are":[3,213],"increasingly":[4],"used":[5],"to":[6,74,87,224],"judge":[7],"people,":[8],"and":[9,68,176],"their":[10],"answers":[11],"can":[12],"depend":[13],"on":[14,112,168],"inputs":[15,41],"correlated":[16],"with":[17,61,95,204],"a":[18,49,69,91,148,214],"protected":[19],"attribute":[20,27],"such":[21,207],"as":[22,125,208],"race,":[23],"even":[24],"when":[25],"that":[26,57,114,150,173],"is":[28,183],"never":[29],"shown.":[30],"Interpreting":[31],"this":[32],"reliance":[33,60,129,144,182,191,228],"requires":[34],"knowing":[35],"how":[36,226],"much":[37],"predictive":[38,106,195],"information":[39],"those":[40],"add,":[42],"but":[43],"audits":[44],"often":[45],"report":[46,190],"only":[47],"whether":[48,180],"decision":[50,174],"changes.":[51],"We":[52,84,99],"introduce":[53],"an":[54,193],"evidence-referenced":[55],"audit":[56],"compares":[58],"proxy":[59,181],"two":[62],"references,":[63],"the":[64,75,108,119,134,186,231],"known":[65,96],"outcome":[66],"rule":[67],"Bayesian":[70],"linear":[71],"regression":[72],"fitted":[73],"same":[76],"labeled":[77],"examples,":[78],"while":[79,118],"measuring":[80],"ranking":[81,157],"accuracy":[82,158,177],"separately.":[83],"apply":[85],"it":[86,222,236],"four":[88,101,154],"in":[90,152,165],"simulated":[92],"patient-ranking":[93],"task":[94],"field":[97],"weights.":[98],"find":[100],"results.":[102],"First,":[103],"at":[104],"zero":[105],"value,":[107],"still":[110],"relied":[111],"fields":[113,200],"carried":[115],"no":[116],"information,":[117],"same-example":[120],"learner":[121],"did":[122,159],"not.":[123],"Second,":[124],"proxies":[126],"became":[127],"predictive,":[128],"grew":[130],"far":[131,227],"less":[132],"than":[133,145,234],"reference.":[135],"Third,":[136],"display":[137],"format":[138],"mattered.":[139],"Ordered":[140,199],"labels":[141],"drew":[142],"more":[143],"arbitrary":[146],"labels,":[147],"direction":[149],"held":[151],"all":[153],"models.":[155],"Fourth,":[156],"not":[160],"consistently":[161],"track":[162],"these":[163,169,219],"changes":[164,175],"reliance.":[166],"Based":[167],"findings,":[170],"we":[171],"argue":[172],"cannot":[178],"show":[179],"warranted":[184],"by":[185],"evidence.":[187],"Audits":[188],"should":[189],"against":[192],"explicit":[194],"reference,":[196],"alongside":[197],"accuracy.":[198],"whose":[201],"values":[202],"move":[203],"group":[205],"membership,":[206],"scores":[209],"or":[210],"income":[211],"bands,":[212],"useful":[215],"starting":[216],"point.":[217],"Together,":[218],"results":[220],"make":[221],"possible":[223],"measure":[225],"departs":[229],"from":[230,237],"evidence,":[232],"rather":[233],"infer":[235],"observed":[238],"choices":[239],"alone.":[240]},"counts_by_year":[],"updated_date":"2026-11-08T12:12:28.789561","created_date":"2026-08-26T00:00:00"}