{"id":"https://openalex.org/W7140125503","doi":"https://doi.org/10.18653/v1/2026.findings-eacl.107","title":"Do Large Language Models Reflect Demographic Pluralism in Safety?","display_name":"Do Large Language Models Reflect Demographic Pluralism in Safety?","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7140125503","doi":"https://doi.org/10.18653/v1/2026.findings-eacl.107"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2026.findings-eacl.107","doi":"https://doi.org/10.18653/v1/2026.findings-eacl.107","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-eacl.107","pdf_url":"https://aclanthology.org/2026.findings-eacl.107.pdf","source":{"id":"https://openalex.org/S7407087051","display_name":"Findings of the Association for Computational Linguistics: EACL","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":"Findings of the Association for Computational Linguistics: EACL 2026","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://aclanthology.org/2026.findings-eacl.107.pdf","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5126968859","display_name":"Usman Naseem","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Usman Naseem","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5076626481","display_name":"Gautam Siddharth Kashyap","orcid":"https://orcid.org/0000-0003-2140-9617","observed_orcids":["https://orcid.org/0000-0003-2140-9617"]},"institutions":[{"id":"https://openalex.org/I99043593","display_name":"Macquarie University","ror":"https://ror.org/01sf06y89","country_code":"AU","type":"education","lineage":["https://openalex.org/I99043593"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Gautam Siddharth Kashyap","raw_affiliation_strings":["Macquarie University , Sydney , Australia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Macquarie University , Sydney , Australia","institution_ids":["https://openalex.org/I99043593"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123504691","display_name":"Sushant Kumar Ray","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I110166357","display_name":"University of Delhi","ror":"https://ror.org/04gzb2213","country_code":"IN","type":"education","lineage":["https://openalex.org/I110166357"]}],"countries":["IN"],"is_corresponding":false,"raw_author_name":"Sushant Kumar Ray","raw_affiliation_strings":["University of Delhi , New Delhi , India"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Delhi , New Delhi , India","institution_ids":["https://openalex.org/I110166357"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5068692517","display_name":"Rafiq Ali","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Rafiq Ali","raw_affiliation_strings":[","],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":",","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5119042124","display_name":"Ebad Shabbir","orcid":"https://orcid.org/0009-0003-0062-9159","observed_orcids":["https://orcid.org/0009-0003-0062-9159"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ebad Shabbir","raw_affiliation_strings":[", 6 DSEU-Okhla , New Delhi , India"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":", 6 DSEU-Okhla , New Delhi , India","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5139254268","display_name":"Abdullah Mohammad","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Abdullah Mohammad","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":19.3287,"has_fulltext":true,"cited_by_count":2,"citation_normalized_percentile":{"value":0.98722102,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":96,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"2042","last_page":"2052"},"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":0.8467000126838684,"subfield":{"id":"https://openalex.org/subfields/3311","display_name":"Safety Research"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},"topics":[{"id":"https://openalex.org/T10883","display_name":"Ethics and Social Impacts of AI","score":0.8467000126838684,"subfield":{"id":"https://openalex.org/subfields/3311","display_name":"Safety Research"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T12262","display_name":"Hate Speech and Cyberbullying Detection","score":0.10779999941587448,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.014499999582767487,"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/large-language-model-safety","display_name":"large language model safety","score":0.9480000138282776},{"id":"https://openalex.org/keywords/perceived-safety","display_name":"perceived safety","score":0.5550000071525574},{"id":"https://openalex.org/keywords/safety-domain","display_name":"safety domain","score":0.5009999871253967},{"id":"https://openalex.org/keywords/safety-assessment","display_name":"safety assessment","score":0.4300000071525574}],"concepts":[{"id":"https://openalex.org/C111472728","wikidata":"https://www.wikidata.org/wiki/Q9471","display_name":"Epistemology","level":1,"score":0.527999997138977},{"id":"https://openalex.org/C144024400","wikidata":"https://www.wikidata.org/wiki/Q21201","display_name":"Sociology","level":0,"score":0.4683000147342682},{"id":"https://openalex.org/C49831778","wikidata":"https://www.wikidata.org/wiki/Q879021","display_name":"Pluralism (philosophy)","level":2,"score":0.45210000872612},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.3449999988079071},{"id":"https://openalex.org/C77805123","wikidata":"https://www.wikidata.org/wiki/Q161272","display_name":"Social psychology","level":1,"score":0.2897999882698059},{"id":"https://openalex.org/C118084267","wikidata":"https://www.wikidata.org/wiki/Q26110","display_name":"Positive economics","level":1,"score":0.28049999475479126},{"id":"https://openalex.org/C9652623","wikidata":"https://www.wikidata.org/wiki/Q190109","display_name":"Field (mathematics)","level":2,"score":0.2793000042438507},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.2687000036239624},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.2606000006198883},{"id":"https://openalex.org/C2779313563","wikidata":"https://www.wikidata.org/wiki/Q17072565","display_name":"On Language","level":2,"score":0.25110000371932983}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.findings-eacl.107","doi":"https://doi.org/10.18653/v1/2026.findings-eacl.107","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-eacl.107","pdf_url":"https://aclanthology.org/2026.findings-eacl.107.pdf","source":{"id":"https://openalex.org/S7407087051","display_name":"Findings of the Association for Computational Linguistics: EACL","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":"Findings of the Association for Computational Linguistics: EACL 2026","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-eacl.107","doi":"https://doi.org/10.18653/v1/2026.findings-eacl.107","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-eacl.107","pdf_url":"https://aclanthology.org/2026.findings-eacl.107.pdf","source":{"id":"https://openalex.org/S7407087051","display_name":"Findings of the Association for Computational Linguistics: EACL","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":"Findings of the Association for Computational Linguistics: EACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/10","score":0.7578999996185303,"display_name":"Reduced inequalities"}],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7140125503.pdf","grobid_xml":"https://content.openalex.org/works/W7140125503.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7106207348","https://openalex.org/W7120272423","https://openalex.org/W7163187240","https://openalex.org/W7155654058","https://openalex.org/W7160597081","https://openalex.org/W4409361921","https://openalex.org/W4386755570","https://openalex.org/W4394867286","https://openalex.org/W4402670259","https://openalex.org/W4414858165"],"abstract_inverted_index":{"Large":[0],"Language":[1],"Model":[2],"(LLM)":[3],"safety":[4,35,66,120],"is":[5,91],"inherently":[6],"pluralistic,":[7],"reflecting":[8],"variations":[9],"in":[10,34],"moral":[11],"norms,":[12],"cultural":[13],"expectations,":[14],"and":[15,24,76,96,110,126],"demographic":[16,45,74,112],"contexts.Yet,":[17],"existing":[18],"alignment":[19],"datasets":[20],"such":[21],"as":[22],"ANTHROPIC-HH":[23],"DICES":[25,61],"rely":[26],"on":[27],"demographically":[28,127],"narrow":[29],"annotator":[30],"pools,":[31],"overlooking":[32],"variation":[33],"perception":[36],"across":[37],"communities.Demo-SafetyBench":[38],"1":[39],"addresses":[40],"this":[41],"gap":[42],"by":[43],"modeling":[44],"pluralism":[46],"directly":[47],"at":[48],"the":[49],"prompt":[50],"level,":[51],"decoupling":[52],"value":[53],"framing":[54],"from":[55,60,69],"responses.In":[56],"Stage":[57,87],"I,":[58],"prompts":[59],"are":[62],"reclassified":[63],"into":[64],"14":[65],"domains":[67,79],"(adapted":[68],"BEAVER-TAILS)":[70],"using":[71,93],"Mistral-7B-Instruct-v0.3,":[72],"retaining":[73],"metadata":[75],"expanding":[77],"low-resource":[78],"via":[80],"Llama-3.1-8B-Instructwith":[81],"SimHash-based":[82],"deduplication,":[83],"yielding":[84],"43,050":[85],"samples.In":[86],"II,":[88],"pluralistic":[89,119],"sensitivity":[90,113],"evaluated":[92],"LLMs-as-Raters-Gemma-7B,":[94],"GPT-4o,":[95],"LLaMA-2-7B-under":[97],"zero-shot":[98],"inference.Balanced":[99],"thresholds":[100],"(\u03b4=0.5,":[101],"\u03c4":[102],"=10)":[103],"achieve":[104],"high":[105],"reliability":[106],"(ICC":[107],"=":[108,115],"0.87)":[109],"low":[111],"(DS":[114],"0.12),":[116],"confirming":[117],"that":[118],"evaluation":[121],"can":[122],"be":[123],"both":[124],"scalable":[125],"robust.":[128]},"counts_by_year":[{"year":2026,"cited_by_count":2}],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-02-12T00:00:00"}