{"id":"https://openalex.org/W4412888805","doi":"https://doi.org/10.18653/v1/2025.findings-acl.145","title":"Cross-Lingual Transfer of Debiasing and Detoxification in Multilingual LLMs: An Extensive Investigation","display_name":"Cross-Lingual Transfer of Debiasing and Detoxification in Multilingual LLMs: An Extensive Investigation","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412888805","doi":"https://doi.org/10.18653/v1/2025.findings-acl.145"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.findings-acl.145","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.145","pdf_url":"https://aclanthology.org/2025.findings-acl.145.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 2025","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.findings-acl.145.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5099123310","display_name":"Vera Neplenbroek","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Vera Neplenbroek","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5019968969","display_name":"Arianna Bisazza","orcid":"https://orcid.org/0000-0003-1270-3048"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Arianna Bisazza","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5103024385","display_name":"Raquel Fern\u00e1ndez","orcid":"https://orcid.org/0000-0002-3889-6194"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Raquel Fern\u00e1ndez","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":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"2805","last_page":"2830"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.9958000183105469,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.9958000183105469,"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/T12881","display_name":"linguistics and terminology studies","score":0.9348999857902527,"subfield":{"id":"https://openalex.org/subfields/1203","display_name":"Language and Linguistics"},"field":{"id":"https://openalex.org/fields/12","display_name":"Arts and Humanities"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T13629","display_name":"Text Readability and Simplification","score":0.9047999978065491,"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":0.9804127216339111},{"id":"https://openalex.org/keywords/detoxification","display_name":"Detoxification (alternative medicine)","score":0.5386284589767456},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.4994642734527588},{"id":"https://openalex.org/keywords/psychology","display_name":"Psychology","score":0.14383557438850403},{"id":"https://openalex.org/keywords/medicine","display_name":"Medicine","score":0.0858151912689209},{"id":"https://openalex.org/keywords/cognitive-science","display_name":"Cognitive science","score":0.07804900407791138}],"concepts":[{"id":"https://openalex.org/C2779458634","wikidata":"https://www.wikidata.org/wiki/Q24963715","display_name":"Debiasing","level":2,"score":0.9804127216339111},{"id":"https://openalex.org/C2780497538","wikidata":"https://www.wikidata.org/wiki/Q1192006","display_name":"Detoxification (alternative medicine)","level":3,"score":0.5386284589767456},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4994642734527588},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.14383557438850403},{"id":"https://openalex.org/C71924100","wikidata":"https://www.wikidata.org/wiki/Q11190","display_name":"Medicine","level":0,"score":0.0858151912689209},{"id":"https://openalex.org/C188147891","wikidata":"https://www.wikidata.org/wiki/Q147638","display_name":"Cognitive science","level":1,"score":0.07804900407791138},{"id":"https://openalex.org/C204787440","wikidata":"https://www.wikidata.org/wiki/Q188504","display_name":"Alternative medicine","level":2,"score":0.0},{"id":"https://openalex.org/C142724271","wikidata":"https://www.wikidata.org/wiki/Q7208","display_name":"Pathology","level":1,"score":0.0}],"mesh":[],"locations_count":4,"locations":[{"id":"doi:10.18653/v1/2025.findings-acl.145","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.145","pdf_url":"https://aclanthology.org/2025.findings-acl.145.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 2025","raw_type":"proceedings-article"},{"id":"pmh:oai:dare.uva.nl:openaire/ad9b5c02-433f-4f83-8c96-288c1559fdbc","is_oa":true,"landing_page_url":"https://handle.uba.uva.nl/personal/pure/en/publications/crosslingual-transfer-of-debiasing-and-detoxification-in-multilingual-llms-an-extensive-investigation(ad9b5c02-433f-4f83-8c96-288c1559fdbc).html","pdf_url":null,"source":{"id":"https://openalex.org/S4306400088","display_name":"UvA-DARE (University of Amsterdam)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I887064364","host_organization_name":"University of Amsterdam","host_organization_lineage":["https://openalex.org/I887064364"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":null,"raw_type":"info:eu-repo/semantics/conferenceObject"},{"id":"pmh:oai:pure.rug.nl:openaire_cris_publications/01449886-3816-47aa-802e-966074391b5e","is_oa":true,"landing_page_url":"https://research.rug.nl/en/publications/01449886-3816-47aa-802e-966074391b5e","pdf_url":null,"source":{"id":"https://openalex.org/S4306400420","display_name":"University of Groningen research database (University of Groningen / Centre for Information Technology)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I169381384","host_organization_name":"University of Groningen","host_organization_lineage":["https://openalex.org/I169381384"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":null,"raw_type":"info:eu-repo/semantics/conferenceObject"},{"id":"pmh:oai:pure.rug.nl:openaire/01449886-3816-47aa-802e-966074391b5e","is_oa":true,"landing_page_url":"https://hdl.handle.net/11370/01449886-3816-47aa-802e-966074391b5e","pdf_url":null,"source":{"id":"https://openalex.org/S4306400420","display_name":"University of Groningen research database (University of Groningen / Centre for Information Technology)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I169381384","host_organization_name":"University of Groningen","host_organization_lineage":["https://openalex.org/I169381384"],"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":"info:eu-repo/semantics/conferenceObject"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-acl.145","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.145","pdf_url":"https://aclanthology.org/2025.findings-acl.145.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 2025","raw_type":"proceedings-article"},"sustainable_development_goals":[{"score":0.6100000143051147,"id":"https://metadata.un.org/sdg/4","display_name":"Quality Education"}],"awards":[{"id":"https://openalex.org/G2388848714","display_name":null,"funder_award_id":"819455","funder_id":"https://openalex.org/F4320320300","funder_display_name":"European Commission"},{"id":"https://openalex.org/G4352045484","display_name":"Polyglot Machines: Human-like Learning of Morphologically Rich Languages","funder_award_id":"VI.Vidi.221C.009","funder_id":"https://openalex.org/F4320321800","funder_display_name":"Nederlandse Organisatie voor Wetenschappelijk Onderzoek"},{"id":"https://openalex.org/G6280615682","display_name":"Low Resource Chat-based Conversational Intelligence","funder_award_id":"NWA.1389.20.183","funder_id":"https://openalex.org/F4320321800","funder_display_name":"Nederlandse Organisatie voor Wetenschappelijk Onderzoek"}],"funders":[{"id":"https://openalex.org/F4320320300","display_name":"European Commission","ror":"https://ror.org/00k4n6c32"},{"id":"https://openalex.org/F4320321800","display_name":"Nederlandse Organisatie voor Wetenschappelijk Onderzoek","ror":"https://ror.org/04jsz6e67"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412888805.pdf","grobid_xml":"https://content.openalex.org/works/W4412888805.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W4362554880","https://openalex.org/W4281684980","https://openalex.org/W4386875279","https://openalex.org/W2171721708","https://openalex.org/W4390963114","https://openalex.org/W4287887864","https://openalex.org/W3214527415"],"abstract_inverted_index":{"Recent":[0],"generative":[1],"large":[2],"language":[3,132,153],"models":[4],"(LLMs)":[5],"show":[6],"remarkable":[7],"performance":[8],"in":[9,15,44,104,129,134,156],"non-English":[10,109,157],"languages,":[11,158],"but":[12,67,86],"when":[13],"prompted":[14],"those":[16],"languages":[17],"they":[18],"tend":[19],"to":[20,48,72,92,108,116],"express":[21],"higher":[22],"harmful":[23],"social":[24],"biases":[25,79],"and":[26,41,65,75,143,166],"toxicity":[27,144,167],"levels.Prior":[28],"work":[29],"has":[30],"shown":[31],"that":[32,113],"finetuning":[33,59,81],"on":[34,61,69,82],"specialized":[35],"datasets":[36],"can":[37,46,121],"mitigate":[38],"this":[39,51,139],"behavior,":[40],"doing":[42],"so":[43],"English":[45,105],"transfer":[47,118,140],"other":[49],"languages.In":[50],"work,":[52],"we":[53],"investigate":[54],"the":[55,62,114,125,135,149,160],"impact":[56],"of":[57,127,141,151,162],"different":[58],"methods":[60,103],"model's":[63,136],"bias":[64,142,165],"toxicity,":[66],"also":[68,106],"its":[70],"ability":[71,155],"produce":[73],"fluent":[74],"diverse":[76],"text.We":[77],"reduce":[78],"by":[80,100,124],"curated":[83],"non-harmful":[84],"text,":[85],"find":[87,111],"only":[88],"direct":[89],"preference":[90],"optimization":[91],"be":[93,122],"effective":[94],"for":[95],"mitigating":[96],"toxicity.The":[97],"mitigation":[98,145,168],"caused":[99],"applying":[101],"these":[102],"transfers":[107],"languages.We":[110],"evidence":[112],"extent":[115],"which":[117],"takes":[119],"place":[120],"predicted":[123],"amount":[126],"data":[128],"a":[130],"given":[131],"present":[133],"pretraining":[137],"data.However,":[138],"often":[146],"comes":[147],"at":[148],"expense":[150],"decreased":[152],"generation":[154],"highlighting":[159],"importance":[161],"developing":[163],"language-specific":[164],"methods.":[169],"1":[170]},"counts_by_year":[],"updated_date":"2026-07-20T07:56:41.581041","created_date":"2025-10-10T00:00:00"}