{"id":"https://openalex.org/W4411523116","doi":"https://doi.org/10.1145/3728881","title":"Fairness Mediator: Neutralize Stereotype Associations to Mitigate Bias in Large Language Models","display_name":"Fairness Mediator: Neutralize Stereotype Associations to Mitigate Bias in Large Language Models","publication_year":2025,"publication_date":"2025-06-22","ids":{"openalex":"https://openalex.org/W4411523116","doi":"https://doi.org/10.1145/3728881"},"language":"en","primary_location":{"id":"doi:10.1145/3728881","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3728881","pdf_url":null,"source":{"id":"https://openalex.org/S4404663975","display_name":"Proceedings of the ACM on software engineering.","issn_l":"2994-970X","issn":["2994-970X"],"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the ACM on Software Engineering","raw_type":"journal-article"},"type":"article","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://doi.org/10.1145/3728881","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5076884053","display_name":"Yisong Xiao","orcid":"https://orcid.org/0000-0001-8227-0052"},"institutions":[{"id":"https://openalex.org/I82880672","display_name":"Beihang University","ror":"https://ror.org/00wk2mp56","country_code":"CN","type":"education","lineage":["https://openalex.org/I82880672"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yisong Xiao","raw_affiliation_strings":["Beihang University, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0001-8227-0052","affiliations":[{"raw_affiliation_string":"Beihang University, Beijing, China","institution_ids":["https://openalex.org/I82880672"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5014870180","display_name":"Aishan Liu","orcid":"https://orcid.org/0000-0002-4224-1318"},"institutions":[{"id":"https://openalex.org/I82880672","display_name":"Beihang University","ror":"https://ror.org/00wk2mp56","country_code":"CN","type":"education","lineage":["https://openalex.org/I82880672"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Aishan Liu","raw_affiliation_strings":["Beihang University, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-4224-1318","affiliations":[{"raw_affiliation_string":"Beihang University, Beijing, China","institution_ids":["https://openalex.org/I82880672"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5081392445","display_name":"Siyuan Liang","orcid":"https://orcid.org/0000-0002-6154-0233"},"institutions":[{"id":"https://openalex.org/I165932596","display_name":"National University of Singapore","ror":"https://ror.org/01tgyzw49","country_code":"SG","type":"education","lineage":["https://openalex.org/I165932596"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Siyuan Liang","raw_affiliation_strings":["National University of Singapore, Singapore, Singapore"],"raw_orcid":"https://orcid.org/0000-0002-6154-0233","affiliations":[{"raw_affiliation_string":"National University of Singapore, Singapore, Singapore","institution_ids":["https://openalex.org/I165932596"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Xianglong Liu","orcid":"https://orcid.org/0000-0002-7618-3275"},"institutions":[{"id":"https://openalex.org/I82880672","display_name":"Beihang University","ror":"https://ror.org/00wk2mp56","country_code":"CN","type":"education","lineage":["https://openalex.org/I82880672"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xianglong Liu","raw_affiliation_strings":["Beihang University, Beijing, China","Zhongguancun Laboratory, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-7618-3275","affiliations":[{"raw_affiliation_string":"Beihang University, Beijing, China","institution_ids":["https://openalex.org/I82880672"]},{"raw_affiliation_string":"Zhongguancun Laboratory, Beijing, China","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5074103823","display_name":"Dacheng Tao","orcid":"https://orcid.org/0000-0001-7225-5449"},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Dacheng Tao","raw_affiliation_strings":["Nanyang Technological University, Singapore, Singapore"],"raw_orcid":"https://orcid.org/0000-0001-7225-5449","affiliations":[{"raw_affiliation_string":"Nanyang Technological University, Singapore, Singapore","institution_ids":["https://openalex.org/I172675005"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":2.792,"has_fulltext":true,"cited_by_count":2,"citation_normalized_percentile":{"value":0.90874591,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":95,"max":96},"biblio":{"volume":"2","issue":"ISSTA","first_page":"250","last_page":"273"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9970999956130981,"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":0.9970999956130981,"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":0.9815000295639038,"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/T10260","display_name":"Software Engineering Research","score":0.9786999821662903,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/debiasing","display_name":"Debiasing","score":0.9483233690261841},{"id":"https://openalex.org/keywords/stereotype","display_name":"Stereotype (UML)","score":0.6560688614845276},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6266738772392273},{"id":"https://openalex.org/keywords/association","display_name":"Association (psychology)","score":0.5913271307945251},{"id":"https://openalex.org/keywords/adversarial-system","display_name":"Adversarial system","score":0.5162618160247803},{"id":"https://openalex.org/keywords/psychology","display_name":"Psychology","score":0.4912227988243103},{"id":"https://openalex.org/keywords/social-psychology","display_name":"Social psychology","score":0.48954400420188904},{"id":"https://openalex.org/keywords/cognitive-psychology","display_name":"Cognitive psychology","score":0.4796350598335266},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.45844411849975586},{"id":"https://openalex.org/keywords/spurious-relationship","display_name":"Spurious relationship","score":0.4347926676273346},{"id":"https://openalex.org/keywords/value","display_name":"Value (mathematics)","score":0.4301244616508484},{"id":"https://openalex.org/keywords/social-cognition","display_name":"Social cognition","score":0.4297163486480713},{"id":"https://openalex.org/keywords/cognitive-bias","display_name":"Cognitive bias","score":0.4270055890083313},{"id":"https://openalex.org/keywords/cognition","display_name":"Cognition","score":0.39195606112480164},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.34634265303611755},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.2628527283668518}],"concepts":[{"id":"https://openalex.org/C2779458634","wikidata":"https://www.wikidata.org/wiki/Q24963715","display_name":"Debiasing","level":2,"score":0.9483233690261841},{"id":"https://openalex.org/C168127410","wikidata":"https://www.wikidata.org/wiki/Q1754331","display_name":"Stereotype (UML)","level":2,"score":0.6560688614845276},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6266738772392273},{"id":"https://openalex.org/C142853389","wikidata":"https://www.wikidata.org/wiki/Q744778","display_name":"Association (psychology)","level":2,"score":0.5913271307945251},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.5162618160247803},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.4912227988243103},{"id":"https://openalex.org/C77805123","wikidata":"https://www.wikidata.org/wiki/Q161272","display_name":"Social psychology","level":1,"score":0.48954400420188904},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.4796350598335266},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.45844411849975586},{"id":"https://openalex.org/C97256817","wikidata":"https://www.wikidata.org/wiki/Q1462316","display_name":"Spurious relationship","level":2,"score":0.4347926676273346},{"id":"https://openalex.org/C2776291640","wikidata":"https://www.wikidata.org/wiki/Q2912517","display_name":"Value (mathematics)","level":2,"score":0.4301244616508484},{"id":"https://openalex.org/C86658582","wikidata":"https://www.wikidata.org/wiki/Q1432778","display_name":"Social cognition","level":3,"score":0.4297163486480713},{"id":"https://openalex.org/C189216375","wikidata":"https://www.wikidata.org/wiki/Q1127759","display_name":"Cognitive bias","level":3,"score":0.4270055890083313},{"id":"https://openalex.org/C169900460","wikidata":"https://www.wikidata.org/wiki/Q2200417","display_name":"Cognition","level":2,"score":0.39195606112480164},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.34634265303611755},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2628527283668518},{"id":"https://openalex.org/C542102704","wikidata":"https://www.wikidata.org/wiki/Q183257","display_name":"Psychotherapist","level":1,"score":0.0},{"id":"https://openalex.org/C169760540","wikidata":"https://www.wikidata.org/wiki/Q207011","display_name":"Neuroscience","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/3728881","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3728881","pdf_url":null,"source":{"id":"https://openalex.org/S4404663975","display_name":"Proceedings of the ACM on software engineering.","issn_l":"2994-970X","issn":["2994-970X"],"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the ACM on Software Engineering","raw_type":"journal-article"},{"id":"pmh:oai:arXiv.org:2504.07787","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2504.07787","pdf_url":"https://arxiv.org/pdf/2504.07787","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"}],"best_oa_location":{"id":"doi:10.1145/3728881","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3728881","pdf_url":null,"source":{"id":"https://openalex.org/S4404663975","display_name":"Proceedings of the ACM on software engineering.","issn_l":"2994-970X","issn":["2994-970X"],"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the ACM on Software Engineering","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G489730147","display_name":null,"funder_award_id":"62206009","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320320671","display_name":"National Research Foundation","ror":"https://ror.org/05s0g1g46"},{"id":"https://openalex.org/F4320320709","display_name":"National Research Foundation Singapore","ror":"https://ror.org/03cpyc314"},{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320335787","display_name":"Fundamental Research Funds for the Central Universities","ror":null}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":60,"referenced_works":["https://openalex.org/W2109308568","https://openalex.org/W2129217160","https://openalex.org/W2186046013","https://openalex.org/W2623399293","https://openalex.org/W2730550703","https://openalex.org/W2751749598","https://openalex.org/W2792317186","https://openalex.org/W2963526187","https://openalex.org/W3016970897","https://openalex.org/W3034115845","https://openalex.org/W3035241006","https://openalex.org/W3047533667","https://openalex.org/W3103934428","https://openalex.org/W3105882417","https://openalex.org/W3123340107","https://openalex.org/W3128630643","https://openalex.org/W3134354193","https://openalex.org/W3135734416","https://openalex.org/W3168398407","https://openalex.org/W3176477796","https://openalex.org/W3184144760","https://openalex.org/W3185095134","https://openalex.org/W3185212449","https://openalex.org/W3206487987","https://openalex.org/W4205221547","https://openalex.org/W4205460703","https://openalex.org/W4220659214","https://openalex.org/W4224249875","https://openalex.org/W4281657280","https://openalex.org/W4284681038","https://openalex.org/W4284697101","https://openalex.org/W4284709622","https://openalex.org/W4284968624","https://openalex.org/W4287333591","https://openalex.org/W4308641598","https://openalex.org/W4308643061","https://openalex.org/W4308731725","https://openalex.org/W4319736227","https://openalex.org/W4323891867","https://openalex.org/W4384154513","https://openalex.org/W4384304645","https://openalex.org/W4385562710","https://openalex.org/W4386136237","https://openalex.org/W4386242329","https://openalex.org/W4387068389","https://openalex.org/W4388229648","https://openalex.org/W4389209181","https://openalex.org/W4389794056","https://openalex.org/W4391558518","https://openalex.org/W4391579668","https://openalex.org/W4396762170","https://openalex.org/W4398239361","https://openalex.org/W4400267113","https://openalex.org/W4401943272","https://openalex.org/W4402443087","https://openalex.org/W4403780588","https://openalex.org/W4404780679","https://openalex.org/W4411551627","https://openalex.org/W6784167499","https://openalex.org/W6839328737"],"related_works":["https://openalex.org/W3019769704","https://openalex.org/W4287812723","https://openalex.org/W2733554571","https://openalex.org/W2999720361","https://openalex.org/W4385573539","https://openalex.org/W805652236","https://openalex.org/W2586707012","https://openalex.org/W2590474933","https://openalex.org/W4254825027","https://openalex.org/W103936845"],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"(LLMs)":[3],"have":[4,59,73],"demonstrated":[5],"remarkable":[6],"performance":[7],"across":[8,221],"diverse":[9],"applications,":[10],"yet":[11],"they":[12],"inadvertently":[13],"absorb":[14],"spurious":[15],"correlations":[16],"from":[17],"training":[18],"data,":[19],"leading":[20],"to":[21,61,78,134,190,210,241],"stereotype":[22,155,164,175],"associations":[23,32,176],"between":[24],"biased":[25,115,187],"concepts":[26,116,188],"and":[27,34,117,126,148,167],"specific":[28],"social":[29,38,84,118,196,217],"groups.":[30,218],"These":[31],"perpetuate":[33],"even":[35],"amplify":[36],"harmful":[37],"biases,":[39,56],"raising":[40],"significant":[41],"concerns":[42],"about":[43],"fairness,":[44],"which":[45,130],"is":[46,95],"a":[47,99,163],"crucial":[48],"issue":[49],"in":[50,93,108,205,233],"software":[51],"engineering.":[52],"To":[53,138],"mitigate":[54],"such":[55],"prior":[57],"studies":[58],"attempted":[60],"project":[62],"model":[63],"embeddings":[64],"into":[65],"unbiased":[66],"spaces":[67],"during":[68,208],"inference.":[69],"However,":[70],"these":[71],"approaches":[72],"shown":[74],"limited":[75],"effectiveness":[76],"due":[77],"their":[79],"weak":[80],"alignment":[81],"with":[82],"downstream":[83],"biases.":[85],"Inspired":[86],"by":[87,182],"the":[88,109,192,200,212],"observation":[89],"that":[90,114,153,226],"concept":[91],"cognition":[92],"LLMs":[94],"primarily":[96],"represented":[97],"through":[98],"linear":[100],"associative":[101],"memory":[102],"mechanism,":[103],"where":[104],"key-value":[105],"mapping":[106],"occurs":[107],"MLP":[110,179,206],"layers,":[111],"we":[112,141],"posited":[113],"groups":[119,197],"are":[120],"similarly":[121],"encoded":[122,177],"as":[123],"entity":[124],"(key)":[125],"information":[127],"(value)":[128],"pairs,":[129],"can":[131],"be":[132],"manipulated":[133],"promote":[135],"fairer":[136],"associations.":[137,156],"this":[139],"end,":[140],"propose":[142],"Fairness":[143],"Mediator":[144],"(FairMed),":[145],"an":[146,168],"effective":[147],"efficient":[149],"bias":[150,237],"mitigation":[151],"framework":[152,158],"neutralizes":[154],"Our":[157],"comprises":[159],"two":[160],"main":[161],"components:":[162],"association":[165,213],"prober":[166,173],"adversarial":[169,201],"debiasing":[170,202],"neutralizer.":[171],"The":[172],"captures":[174],"within":[178],"layer":[180],"activations":[181,207],"employing":[183],"prompts":[184],"centered":[185],"around":[186],"(keys)":[189],"detect":[191],"emission":[193],"probabilities":[194,214],"for":[195],"(values).":[198],"Subsequently,":[199],"neutralizer":[203],"intervenes":[204],"inference":[209],"equalize":[211],"among":[215],"different":[216],"Extensive":[219],"experiments":[220],"nine":[222],"protected":[223],"attributes":[224],"show":[225],"our":[227],"FairMed":[228],"significantly":[229],"outperforms":[230],"state-of-the-art":[231],"methods":[232],"effectiveness,":[234],"achieving":[235],"average":[236],"reductions":[238],"of":[239],"up":[240],"84.42":[242]},"counts_by_year":[{"year":2025,"cited_by_count":2}],"updated_date":"2026-05-21T06:26:12.895304","created_date":"2025-10-10T00:00:00"}