{"id":"https://openalex.org/W4416036063","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.21","title":"sudoLLM: On Multi-role Alignment of Language Models","display_name":"sudoLLM: On Multi-role Alignment of Language Models","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4416036063","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.21"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.21","is_oa":false,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.21","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2025","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2505.14607","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5030585541","display_name":"Soumadeep Saha","orcid":"https://orcid.org/0000-0003-0013-5143"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Soumadeep Saha","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5019757348","display_name":"Akshay Chaturvedi","orcid":"https://orcid.org/0000-0002-8192-6221"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Akshay Chaturvedi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5087199286","display_name":"Joy Mahapatra","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Joy Mahapatra","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5024170958","display_name":"Utpal Garain","orcid":"https://orcid.org/0000-0001-7207-5018"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Utpal Garain","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":0.37183014,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"366","last_page":"384"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10927","display_name":"Access Control and Trust","score":0.6959999799728394,"subfield":{"id":"https://openalex.org/subfields/3312","display_name":"Sociology and Political Science"},"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/T10927","display_name":"Access Control and Trust","score":0.6959999799728394,"subfield":{"id":"https://openalex.org/subfields/3312","display_name":"Sociology and Political Science"},"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/T11424","display_name":"Security and Verification in Computing","score":0.03180000185966492,"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/T10734","display_name":"Information and Cyber Security","score":0.020400000736117363,"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/language-model","display_name":"Language model","score":0.6261000037193298},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.4984000027179718},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.47690001130104065},{"id":"https://openalex.org/keywords/access-control","display_name":"Access control","score":0.4717000126838684},{"id":"https://openalex.org/keywords/signal","display_name":"SIGNAL (programming language)","score":0.4000000059604645},{"id":"https://openalex.org/keywords/train","display_name":"Train","score":0.3614000082015991},{"id":"https://openalex.org/keywords/control","display_name":"Control (management)","score":0.3490000069141388}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7490000128746033},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.6261000037193298},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.4984000027179718},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.47690001130104065},{"id":"https://openalex.org/C527821871","wikidata":"https://www.wikidata.org/wiki/Q228502","display_name":"Access control","level":2,"score":0.4717000126838684},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.43650001287460327},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.41339999437332153},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.4000000059604645},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3707999885082245},{"id":"https://openalex.org/C190839683","wikidata":"https://www.wikidata.org/wiki/Q2448197","display_name":"Train","level":2,"score":0.3614000082015991},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.3490000069141388},{"id":"https://openalex.org/C182306322","wikidata":"https://www.wikidata.org/wiki/Q1779371","display_name":"Order (exchange)","level":2,"score":0.3467000126838684},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.3199999928474426},{"id":"https://openalex.org/C2776543384","wikidata":"https://www.wikidata.org/wiki/Q593289","display_name":"Information access","level":2,"score":0.3165999948978424},{"id":"https://openalex.org/C179603123","wikidata":"https://www.wikidata.org/wiki/Q1941921","display_name":"Modeling language","level":3,"score":0.30880001187324524},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.3068999946117401},{"id":"https://openalex.org/C94922259","wikidata":"https://www.wikidata.org/wiki/Q33215","display_name":"Constructed language","level":2,"score":0.2944999933242798},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.289000004529953},{"id":"https://openalex.org/C120936955","wikidata":"https://www.wikidata.org/wiki/Q2155640","display_name":"Empirical research","level":2,"score":0.26980000734329224},{"id":"https://openalex.org/C140547941","wikidata":"https://www.wikidata.org/wiki/Q7797194","display_name":"Threat model","level":2,"score":0.25369998812675476}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.18653/v1/2025.findings-emnlp.21","is_oa":false,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.21","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2025","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2505.14607","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2505.14607","pdf_url":"https://arxiv.org/pdf/2505.14607","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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2505.14607","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2505.14607","pdf_url":"https://arxiv.org/pdf/2505.14607","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"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416036063.pdf","grobid_xml":"https://content.openalex.org/works/W4416036063.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"User":[0],"authorization-based":[1],"access":[2,32,58],"privileges":[3],"are":[4],"a":[5,38],"key":[6],"feature":[7],"in":[8,18,43,54,76],"many":[9],"safety-critical":[10],"systems,":[11,34],"but":[12],"have":[13],"not":[14],"been":[15],"extensively":[16],"studied":[17],"the":[19,86,114,132,135],"large":[20],"language":[21,115],"model":[22],"(LLM)":[23],"realm.":[24],"In":[25],"this":[26,73,96],"work,":[27],"drawing":[28],"inspiration":[29],"from":[30],"such":[31],"control":[33],"we":[35],"introduce":[36],"sudoLLM,":[37],"novel":[39],"framework":[40,140],"that":[41,49,95],"results":[42,93],"multi-role":[44],"aligned":[45],"LLMs,":[46,127],"i.e.,":[47],"LLMs":[48],"account":[50],"for,":[51],"and":[52,67,83,108,118,148],"behave":[53],"accordance":[55],"with,":[56],"user":[57,87],"rights.":[59],"sudoLLM":[60],"injects":[61],"subtle":[62],"user-based":[63],"biases":[64],"into":[65],"queries":[66],"trains":[68],"an":[69,144],"LLM":[70],"to":[71,78,104,125],"utilize":[72],"bias":[74,137],"signal":[75],"order":[77],"produce":[79],"sensitive":[80],"information":[81],"if":[82,85],"only":[84],"is":[88,122,128,141],"authorized.":[89],"We":[90],"present":[91],"empirical":[92],"demonstrating":[94],"approach":[97],"shows":[98],"substantially":[99],"improved":[100],"alignment,":[101,120],"generalization,":[102],"resistance":[103],"prefix-based":[105],"jailbreaking":[106],"attacks,":[107],"``fails-closed''.":[109],"The":[110],"persistent":[111],"tension":[112],"between":[113],"modeling":[116],"objective":[117],"safety":[119,156],"which":[121],"often":[123],"exploited":[124],"jailbreak":[126],"somewhat":[129],"resolved":[130],"with":[131,157],"aid":[133],"of":[134],"injected":[136],"signal.":[138],"Our":[139],"meant":[142],"as":[143],"additional":[145],"security":[146],"layer,":[147],"complements":[149],"existing":[150],"guardrail":[151],"mechanisms":[152],"for":[153],"enhanced":[154],"end-to-end":[155],"LLMs.":[158]},"counts_by_year":[],"updated_date":"2026-08-18T07:49:30.821534","created_date":"2025-11-08T00:00:00"}
