{"id":"https://openalex.org/W4416452319","doi":"https://doi.org/10.18653/v1/2026.findings-acl.971","title":"Large Language Models Are Still Misled by Simple Bias Ensembles","display_name":"Large Language Models Are Still Misled by Simple Bias Ensembles","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W4416452319","doi":"https://doi.org/10.18653/v1/2026.findings-acl.971"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2026.findings-acl.971","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.971","pdf_url":"https://aclanthology.org/2026.findings-acl.971.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.findings-acl.971.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100584977","display_name":"Zhouhao Sun","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhouhao Sun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5120558250","display_name":"Zhiyuan Kan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhiyuan Kan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102305510","display_name":"Xiao Ding","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiao Ding","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5109975421","display_name":"Li Du","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li Du","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Bibo Cai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bibo Cai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Yang Zhao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang Zhao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Bing Qin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bing Qin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5064463814","display_name":"Ting Liu","orcid":"https://orcid.org/0000-0003-3725-6408"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ting Liu","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":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.02625279,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"19444","last_page":"19455"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.30970001220703125,"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":0.30970001220703125,"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":0.23199999332427979,"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/T11303","display_name":"Bayesian Modeling and Causal Inference","score":0.04639999940991402,"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/debiasing","display_name":"Debiasing","score":0.9793000221252441},{"id":"https://openalex.org/keywords/generalizability-theory","display_name":"Generalizability theory","score":0.9128000140190125},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.6593000292778015},{"id":"https://openalex.org/keywords/benchmarking","display_name":"Benchmarking","score":0.6424999833106995},{"id":"https://openalex.org/keywords/causal-model","display_name":"Causal model","score":0.35910001397132874},{"id":"https://openalex.org/keywords/causality","display_name":"Causality (physics)","score":0.3212999999523163},{"id":"https://openalex.org/keywords/bridge","display_name":"Bridge (graph theory)","score":0.31929999589920044}],"concepts":[{"id":"https://openalex.org/C2779458634","wikidata":"https://www.wikidata.org/wiki/Q24963715","display_name":"Debiasing","level":2,"score":0.9793000221252441},{"id":"https://openalex.org/C27158222","wikidata":"https://www.wikidata.org/wiki/Q5532422","display_name":"Generalizability theory","level":2,"score":0.9128000140190125},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.6593000292778015},{"id":"https://openalex.org/C86251818","wikidata":"https://www.wikidata.org/wiki/Q816754","display_name":"Benchmarking","level":2,"score":0.6424999833106995},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6040999889373779},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.4336000084877014},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4242999851703644},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.39969998598098755},{"id":"https://openalex.org/C11671645","wikidata":"https://www.wikidata.org/wiki/Q5054567","display_name":"Causal model","level":2,"score":0.35910001397132874},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.34700000286102295},{"id":"https://openalex.org/C64357122","wikidata":"https://www.wikidata.org/wiki/Q1149766","display_name":"Causality (physics)","level":2,"score":0.3212999999523163},{"id":"https://openalex.org/C100776233","wikidata":"https://www.wikidata.org/wiki/Q2532492","display_name":"Bridge (graph theory)","level":2,"score":0.31929999589920044},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.31310001015663147},{"id":"https://openalex.org/C78022038","wikidata":"https://www.wikidata.org/wiki/Q7644082","display_name":"Superordinate goals","level":2,"score":0.31119999289512634},{"id":"https://openalex.org/C62354387","wikidata":"https://www.wikidata.org/wiki/Q875399","display_name":"Boundary (topology)","level":2,"score":0.30889999866485596},{"id":"https://openalex.org/C2987370644","wikidata":"https://www.wikidata.org/wiki/Q7836903","display_name":"Treatment effect","level":2,"score":0.30660000443458557},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.3003999888896942},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.29440000653266907},{"id":"https://openalex.org/C48372109","wikidata":"https://www.wikidata.org/wiki/Q3913","display_name":"Binary number","level":2,"score":0.25450000166893005},{"id":"https://openalex.org/C2779628075","wikidata":"https://www.wikidata.org/wiki/Q1253258","display_name":"Downgrade","level":2,"score":0.25}],"mesh":[],"locations_count":4,"locations":[{"id":"doi:10.18653/v1/2026.findings-acl.971","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.971","pdf_url":"https://aclanthology.org/2026.findings-acl.971.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2505.16522","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2505.16522","pdf_url":"https://arxiv.org/pdf/2505.16522","source":{"id":"https://openalex.org/S4393918464","display_name":"ArXiv.org","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"pmh:oai:arXiv.org:2505.16522","is_oa":true,"landing_page_url":"https://arxiv.org/abs/2505.16522","pdf_url":"https://arxiv.org/pdf/2505.16522","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,"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2505.16522","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2505.16522","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,"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":"Preprint"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-acl.971","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.971","pdf_url":"https://aclanthology.org/2026.findings-acl.971.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G4442529403","display_name":null,"funder_award_id":"U22B2059","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416452319.pdf","grobid_xml":"https://content.openalex.org/works/W4416452319.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"With":[0],"the":[1,20,115],"evolution":[2],"of":[3,22,45,83,99,117],"large":[4],"language":[5],"models":[6],"(LLMs),":[7],"their":[8],"robustness":[9],"against":[10],"individual":[11],"simple":[12,24],"biases":[13,25],"has":[14],"been":[15],"enhanced.However,":[16],"we":[17,88],"observe":[18],"that":[19,34,103],"ensemble":[21],"multiple":[23,97],"still":[26],"exerts":[27],"a":[28,42,90],"significant":[29],"adverse":[30],"impact":[31],"on":[32,111],"LLMs.Given":[33],"real-world":[35],"data":[36],"samples":[37],"are":[38,69],"typically":[39],"confounded":[40],"by":[41],"wide":[43],"range":[44],"biases,":[46],"LLMs":[47,105],"tend":[48],"to":[49,71],"exhibit":[50],"unstable":[51],"performance":[52],"when":[53],"deployed":[54],"in":[55],"high-stakes":[56],"realworld":[57],"scenarios":[58],"such":[59,119],"as":[60],"clinical":[61],"diagnosis":[62],"and":[63,106],"legal":[64],"document":[65],"analysis.However,":[66],"previous":[67],"benchmarks":[68],"constrained":[70],"datasets":[72],"where":[73,93],"each":[74,94],"sample":[75,95],"is":[76],"manually":[77],"injected":[78],"with":[79],"only":[80],"one":[81],"type":[82],"bias.To":[84],"bridge":[85],"this":[86,112],"gap,":[87],"propose":[89],"multi-bias":[91],"benchmark":[92],"contains":[96],"types":[98],"biases.Experimental":[100],"results":[101],"reveal":[102],"existing":[104],"debiasing":[107],"methods":[108],"perform":[109],"poorly":[110],"benchmark,":[113],"highlighting":[114],"challenge":[116],"eliminating":[118],"compounded":[120],"biases.":[121]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-10-10T00:00:00"}