{"id":"https://openalex.org/W4411549853","doi":"https://doi.org/10.1145/3715275.3732196","title":"The Root Shapes the Fruit: On the Persistence of Gender-Exclusive Harms in Aligned Language Models","display_name":"The Root Shapes the Fruit: On the Persistence of Gender-Exclusive Harms in Aligned Language Models","publication_year":2025,"publication_date":"2025-06-23","ids":{"openalex":"https://openalex.org/W4411549853","doi":"https://doi.org/10.1145/3715275.3732196"},"language":"en","primary_location":{"id":"doi:10.1145/3715275.3732196","doi":"https://doi.org/10.1145/3715275.3732196","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3715275.3732196","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3715275.3732196","source":{"id":"https://openalex.org/S4363608463","display_name":"ACM Conference on Fairness, Accountability, and Transparency (FAccT)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"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 2025 ACM Conference on Fairness, Accountability, and Transparency","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","is_traditional":true,"indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3715275.3732196","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5064234845","display_name":"Anaelia Ovalle","orcid":"https://orcid.org/0000-0002-0531-7520","observed_orcids":["https://orcid.org/0000-0002-0531-7520"]},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Anaelia Ovalle","raw_affiliation_strings":["Meta, New York, USA"],"raw_orcid":"https://orcid.org/0000-0002-0531-7520","affiliations":[{"raw_affiliation_string":"Meta, New York, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5114641229","display_name":"Krunoslav Lehman Pavasovic","orcid":"https://orcid.org/0009-0003-9354-1008","observed_orcids":["https://orcid.org/0009-0003-9354-1008"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Krunoslav Lehman Pavasovic","raw_affiliation_strings":["Meta, Paris, France"],"raw_orcid":"https://orcid.org/0009-0003-9354-1008","affiliations":[{"raw_affiliation_string":"Meta, Paris, France","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Louis Martin","orcid":"https://orcid.org/0009-0001-9268-1581","observed_orcids":["https://orcid.org/0009-0001-9268-1581"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Louis Martin","raw_affiliation_strings":["Meta, Paris, France"],"raw_orcid":"https://orcid.org/0009-0001-9268-1581","affiliations":[{"raw_affiliation_string":"Meta, Paris, France","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5067919401","display_name":"Luke Zettlemoyer","orcid":"https://orcid.org/0009-0008-8296-0764","observed_orcids":["https://orcid.org/0009-0008-8296-0764"]},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Luke Zettlemoyer","raw_affiliation_strings":["Meta, Seattle, USA"],"raw_orcid":"https://orcid.org/0009-0008-8296-0764","affiliations":[{"raw_affiliation_string":"Meta, Seattle, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Eric Michael Smith","orcid":"https://orcid.org/0009-0005-3076-1948","observed_orcids":["https://orcid.org/0009-0005-3076-1948"]},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Eric Michael Smith","raw_affiliation_strings":["Meta, New York, USA"],"raw_orcid":"https://orcid.org/0009-0005-3076-1948","affiliations":[{"raw_affiliation_string":"Meta, New York, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5087096372","display_name":"Kai-Wei Chang","orcid":"https://orcid.org/0000-0001-5365-0072","observed_orcids":["https://orcid.org/0000-0001-5365-0072"]},"institutions":[{"id":"https://openalex.org/I161318765","display_name":"University of California, Los Angeles","ror":"https://ror.org/046rm7j60","country_code":"US","type":"education","lineage":["https://openalex.org/I161318765"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Kai-Wei Chang","raw_affiliation_strings":["UCLA, Los Angeles, USA"],"raw_orcid":"https://orcid.org/0000-0001-5365-0072","affiliations":[{"raw_affiliation_string":"UCLA, Los Angeles, USA","institution_ids":["https://openalex.org/I161318765"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5062696000","display_name":"Adina Williams","orcid":"https://orcid.org/0000-0001-5281-3343","observed_orcids":["https://orcid.org/0000-0001-5281-3343"]},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Adina Williams","raw_affiliation_strings":["Meta, New York, USA"],"raw_orcid":"https://orcid.org/0000-0001-5281-3343","affiliations":[{"raw_affiliation_string":"Meta, New York, USA","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5071931582","display_name":"Levent Sagun","orcid":"https://orcid.org/0000-0001-5403-4124","observed_orcids":["https://orcid.org/0000-0001-5403-4124"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Levent Sagun","raw_affiliation_strings":["Meta, Paris, France"],"raw_orcid":"https://orcid.org/0000-0001-5403-4124","affiliations":[{"raw_affiliation_string":"Meta, Paris, France","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":3.9234,"has_fulltext":true,"cited_by_count":12,"citation_normalized_percentile":{"value":0.94595158,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":88,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"3094","last_page":"3105"},"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":0.9887999892234802,"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":0.9887999892234802,"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/T12262","display_name":"Hate Speech and Cyberbullying Detection","score":0.0031999999191612005,"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/T10028","display_name":"Topic Modeling","score":0.0012000000569969416,"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/language-model-alignment","display_name":"language model alignment","score":0.6389999985694885},{"id":"https://openalex.org/keywords/preference-optimization","display_name":"preference optimization","score":0.5830000042915344},{"id":"https://openalex.org/keywords/supervised-fine-tuning","display_name":"supervised fine-tuning","score":0.6819999814033508},{"id":"https://openalex.org/keywords/stigmatization","display_name":"stigmatization","score":0.9229999780654907},{"id":"https://openalex.org/keywords/underrepresented-groups","display_name":"underrepresented groups","score":0.7269999980926514}],"concepts":[{"id":"https://openalex.org/C2781009140","wikidata":"https://www.wikidata.org/wiki/Q7170389","display_name":"Persistence (discontinuity)","level":2,"score":0.7840172052383423},{"id":"https://openalex.org/C171078966","wikidata":"https://www.wikidata.org/wiki/Q111029","display_name":"Root (linguistics)","level":2,"score":0.7139396667480469},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5258116126060486},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.23209542036056519},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.1333356499671936},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0678713321685791},{"id":"https://openalex.org/C187320778","wikidata":"https://www.wikidata.org/wiki/Q1349130","display_name":"Geotechnical engineering","level":1,"score":0.0}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1145/3715275.3732196","doi":"https://doi.org/10.1145/3715275.3732196","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3715275.3732196","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3715275.3732196","source":{"id":"https://openalex.org/S4363608463","display_name":"ACM Conference on Fairness, Accountability, and Transparency (FAccT)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"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 2025 ACM Conference on Fairness, Accountability, and Transparency","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2411.03700","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2411.03700","pdf_url":"https://arxiv.org/pdf/2411.03700","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.2411.03700","doi":"https://doi.org/10.48550/arxiv.2411.03700","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2411.03700","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":"doi:10.1145/3715275.3732196","doi":"https://doi.org/10.1145/3715275.3732196","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3715275.3732196","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3715275.3732196","source":{"id":"https://openalex.org/S4363608463","display_name":"ACM Conference on Fairness, Accountability, and Transparency (FAccT)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"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 2025 ACM Conference on Fairness, Accountability, and Transparency","raw_type":"proceedings-article"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/10","display_name":"Reduced inequalities","score":0.8974000215530396},{"id":"https://metadata.un.org/sdg/5","display_name":"Gender equality","score":0.6025999784469604}],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4411549853.pdf","grobid_xml":"https://content.openalex.org/works/W4411549853.grobid-xml"},"referenced_works_count":53,"referenced_works":["https://openalex.org/W2013784666","https://openalex.org/W2736601468","https://openalex.org/W2769358515","https://openalex.org/W2795038878","https://openalex.org/W2798280648","https://openalex.org/W2897042519","https://openalex.org/W2899136066","https://openalex.org/W2942908811","https://openalex.org/W2962787423","https://openalex.org/W2963526187","https://openalex.org/W2971307358","https://openalex.org/W2973379954","https://openalex.org/W2990138404","https://openalex.org/W3034656957","https://openalex.org/W3088599783","https://openalex.org/W3089430725","https://openalex.org/W3123340107","https://openalex.org/W3133702157","https://openalex.org/W3172415559","https://openalex.org/W3198255289","https://openalex.org/W3198409578","https://openalex.org/W3201174429","https://openalex.org/W3207316473","https://openalex.org/W3215698779","https://openalex.org/W4223908421","https://openalex.org/W4226112058","https://openalex.org/W4234580964","https://openalex.org/W4285210581","https://openalex.org/W4296413526","https://openalex.org/W4311991106","https://openalex.org/W4377157114","https://openalex.org/W4377823230","https://openalex.org/W4383175701","https://openalex.org/W4384918448","https://openalex.org/W4385570581","https://openalex.org/W4385572155","https://openalex.org/W4385572534","https://openalex.org/W4385574250","https://openalex.org/W4389471473","https://openalex.org/W4392011551","https://openalex.org/W4392427283","https://openalex.org/W4392616695","https://openalex.org/W4393063868","https://openalex.org/W4394654134","https://openalex.org/W4394906585","https://openalex.org/W4401043164","https://openalex.org/W4401043206","https://openalex.org/W4401209579","https://openalex.org/W4402670301","https://openalex.org/W4403662541","https://openalex.org/W4412886932","https://openalex.org/W4412889417","https://openalex.org/W4412945467"],"related_works":["https://openalex.org/W7140437619","https://openalex.org/W7057764459","https://openalex.org/W7214885126","https://openalex.org/W4417300300","https://openalex.org/W4405182693","https://openalex.org/W7130541037","https://openalex.org/W7150950703","https://openalex.org/W4412886776","https://openalex.org/W4405030380","https://openalex.org/W4404978487"],"abstract_inverted_index":{"Natural-language":[0],"assistants":[1],"are":[2,171],"designed":[3],"to":[4,18,88,145,162,174,201,216],"provide":[5],"users":[6],"with":[7,96,198],"helpful":[8],"responses":[9],"while":[10],"avoiding":[11],"harmful":[12,35,97,154],"outputs,":[13],"largely":[14],"achieved":[15],"through":[16],"alignment":[17,27,91,205],"human":[19],"preferences.Yet":[20],"there":[21],"is":[22,44],"limited":[23],"understanding":[24],"of":[25,49,109,126,183,211],"whether":[26],"techniques":[28],"may":[29,179],"inadvertently":[30],"perpetuate":[31],"or":[32],"even":[33],"amplify":[34,180],"biases":[36,72,99,128,156],"inherited":[37],"from":[38],"their":[39,189],"prealigned":[40],"base":[41,190],"models.This":[42],"issue":[43],"compounded":[45],"by":[46],"the":[47,209],"choice":[48],"bias":[50,110,142,213],"evaluation":[51,111,125,214],"benchmarks":[52,143],"in":[53,100,120,157,188,224],"popular":[54,141],"preference-finetuned":[55,115],"models,":[56],"which":[57],"predominantly":[58],"focus":[59],"on":[60],"dominant":[61],"social":[62,164],"categories,":[63],"such":[64],"as":[65],"binary":[66],"gender,":[67],"thereby":[68],"limiting":[69],"insights":[70],"into":[71],"affecting":[73],"underrepresented":[74,222],"groups.To":[75],"address":[76,221],"this":[77],"gap,":[78],"we":[79],"center":[80],"transgender,":[81],"nonbinary,":[82],"and":[83,94,147,178,193,203,220],"other":[84,163],"gender-diverse":[85,127,155,185],"(TGNB)":[86],"identities":[87],"investigate":[89],"how":[90],"procedures":[92],"encode":[93],"interact":[95],"TGNB":[98],"LLMs.Our":[101],"key":[102],"contributions":[103],"include:":[104],"1)":[105],"a":[106,149],"comprehensive":[107],"survey":[108],"modalities":[112],"across":[113,129,133],"leading":[114],"LLMs,":[116],"highlighting":[117],"critical":[118],"gaps":[119],"genderdiverse":[121],"representation,":[122],"2)":[123],"systematic":[124],"16":[130],"publicly-available":[131],"models":[132,170],"Direct":[134],"Preference":[135],"Optimization":[136],"(DPO)":[137],"stages,":[138],"uncovering":[139],"harms":[140,186,223],"fail":[144],"detect,":[146],"3)":[148],"flexible":[150],"framework":[151],"for":[152,208],"measuring":[153],"implicit":[158],"reward":[159],"signals":[160],"applicable":[161],"contexts.Our":[165],"findings":[166],"suggest":[167],"that":[168],"DPOaligned":[169],"particularly":[172],"sensitive":[173],"supervised":[175],"finetuning":[176],"(SFT),":[177],"two":[181],"forms":[182],"real-world":[184],"reflected":[187],"models:":[191],"stigmatization":[192],"gender":[194],"non-affirmative":[195],"language.We":[196],"conclude":[197],"recommendations":[199],"tailored":[200],"DPO":[202],"broader":[204],"practices,":[206],"advocating":[207],"adoption":[210],"community-informed":[212],"frameworks":[215],"more":[217],"effectively":[218],"identify":[219],"LLMs.":[225]},"counts_by_year":[{"year":2026,"cited_by_count":7},{"year":2025,"cited_by_count":4},{"year":2024,"cited_by_count":1}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2025-10-10T00:00:00"}
