{"id":"https://openalex.org/W4412888004","doi":"https://doi.org/10.18653/v1/2025.findings-acl.763","title":"REVS: Unlearning Sensitive Information in Language Models via Rank Editing in the Vocabulary Space","display_name":"REVS: Unlearning Sensitive Information in Language Models via Rank Editing in the Vocabulary Space","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412888004","doi":"https://doi.org/10.18653/v1/2025.findings-acl.763"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.findings-acl.763","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.763","pdf_url":"https://aclanthology.org/2025.findings-acl.763.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.763.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5099137410","display_name":"Tomer Ashuach","orcid":null},"institutions":[{"id":"https://openalex.org/I174306211","display_name":"Technion \u2013 Israel Institute of Technology","ror":"https://ror.org/03qryx823","country_code":"IL","type":"education","lineage":["https://openalex.org/I174306211"]}],"countries":["IL"],"is_corresponding":false,"raw_author_name":"Tomer Ashuach","raw_affiliation_strings":["Technion -Israel Institute of Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Technion -Israel Institute of Technology","institution_ids":["https://openalex.org/I174306211"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5020691628","display_name":"Martin Tutek","orcid":"https://orcid.org/0000-0001-5227-5397"},"institutions":[{"id":"https://openalex.org/I174306211","display_name":"Technion \u2013 Israel Institute of Technology","ror":"https://ror.org/03qryx823","country_code":"IL","type":"education","lineage":["https://openalex.org/I174306211"]}],"countries":["IL"],"is_corresponding":false,"raw_author_name":"Martin Tutek","raw_affiliation_strings":["Technion -Israel Institute of Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Technion -Israel Institute of Technology","institution_ids":["https://openalex.org/I174306211"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5051184573","display_name":"Yonatan Belinkov","orcid":null},"institutions":[{"id":"https://openalex.org/I174306211","display_name":"Technion \u2013 Israel Institute of Technology","ror":"https://ror.org/03qryx823","country_code":"IL","type":"education","lineage":["https://openalex.org/I174306211"]}],"countries":["IL"],"is_corresponding":false,"raw_author_name":"Yonatan Belinkov","raw_affiliation_strings":["Technion -Israel Institute of Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Technion -Israel Institute of Technology","institution_ids":["https://openalex.org/I174306211"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I174306211"],"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":"14774","last_page":"14797"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.9972000122070312,"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":0.9972000122070312,"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":0.9955000281333923,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9501000046730042,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/vocabulary","display_name":"Vocabulary","score":0.7486662864685059},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7362880110740662},{"id":"https://openalex.org/keywords/rank","display_name":"Rank (graph theory)","score":0.727942943572998},{"id":"https://openalex.org/keywords/space","display_name":"Space (punctuation)","score":0.6341487169265747},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.4715428650379181},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.46755778789520264},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.38486745953559875},{"id":"https://openalex.org/keywords/information-retrieval","display_name":"Information retrieval","score":0.35521620512008667},{"id":"https://openalex.org/keywords/linguistics","display_name":"Linguistics","score":0.194448322057724},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.09997901320457458},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.06833356618881226}],"concepts":[{"id":"https://openalex.org/C2777601683","wikidata":"https://www.wikidata.org/wiki/Q6499736","display_name":"Vocabulary","level":2,"score":0.7486662864685059},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7362880110740662},{"id":"https://openalex.org/C164226766","wikidata":"https://www.wikidata.org/wiki/Q7293202","display_name":"Rank (graph theory)","level":2,"score":0.727942943572998},{"id":"https://openalex.org/C2778572836","wikidata":"https://www.wikidata.org/wiki/Q380933","display_name":"Space (punctuation)","level":2,"score":0.6341487169265747},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4715428650379181},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.46755778789520264},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.38486745953559875},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.35521620512008667},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.194448322057724},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.09997901320457458},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.06833356618881226},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0},{"id":"https://openalex.org/C114614502","wikidata":"https://www.wikidata.org/wiki/Q76592","display_name":"Combinatorics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-acl.763","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.763","pdf_url":"https://aclanthology.org/2025.findings-acl.763.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"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-acl.763","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.763","pdf_url":"https://aclanthology.org/2025.findings-acl.763.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":0.7099999785423279,"display_name":"Quality Education","id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[{"id":"https://openalex.org/F4320315389","display_name":"Open Philanthropy Project","ror":"https://ror.org/004d1k391"},{"id":"https://openalex.org/F4320320300","display_name":"European Commission","ror":"https://ror.org/00k4n6c32"},{"id":"https://openalex.org/F4320323139","display_name":"Azrieli Foundation","ror":"https://ror.org/009svts64"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412888004.pdf","grobid_xml":"https://content.openalex.org/works/W4412888004.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W2349784553","https://openalex.org/W3022596247","https://openalex.org/W2601444686","https://openalex.org/W4307058054","https://openalex.org/W4292238148","https://openalex.org/W4323660495","https://openalex.org/W2385319785","https://openalex.org/W2900827440","https://openalex.org/W3167549738","https://openalex.org/W2381983017"],"abstract_inverted_index":{"Language":[0],"models":[1,106],"(LMs)":[2],"risk":[3],"inadvertently":[4],"memorizing":[5],"and":[6,35,59,87,95,120],"divulging":[7],"sensitive":[8,53,72,80,118],"or":[9,30],"personally":[10],"identifiable":[11],"information":[12,54,119],"(PII)":[13],"seen":[14],"in":[15,116],"training":[16],"data,":[17],"causing":[18],"privacy":[19],"concerns.Current":[20],"approaches":[21],"to":[22,107,109,122],"address":[23],"this":[24],"issue":[25],"involve":[26],"costly":[27],"dataset":[28,101],"scrubbing,":[29],"model":[31,36,128],"filtering":[32],"through":[33,42],"unlearning":[34,52,117],"editing,":[37],"which":[38],"can":[39],"be":[40],"bypassed":[41],"extraction":[43,123],"attacks.We":[44],"propose":[45],"REVS,":[46],"a":[47,61,96],"novel":[48],"non-gradient-based":[49],"method":[50,77],"for":[51,67],"from":[55],"LMs.":[56],"REVS":[57,112],"identifies":[58],"modifies":[60],"small":[62],"subset":[63],"of":[64],"neurons":[65],"relevant":[66],"constituent":[68],"tokens":[69],"that":[70,102],"form":[71],"information.To":[73],"adequately":[74],"evaluate":[75],"our":[76],"on":[78],"truly":[79],"information,":[81],"we":[82,103],"curate":[83],"three":[84],"datasets:":[85],"email":[86],"URL":[88],"datasets":[89],"naturally":[90],"memorized":[91],"by":[92],"the":[93,105],"models,":[94],"synthetic":[97],"social":[98],"security":[99],"number":[100],"tune":[104],"memorize.Compared":[108],"other":[110],"methods,":[111],"demonstrates":[113],"superior":[114],"performance":[115],"robustness":[121],"attacks,":[124],"while":[125],"retaining":[126],"underlying":[127],"integrity.":[129]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}