{"id":"https://openalex.org/W4411120109","doi":"https://doi.org/10.18653/v1/2025.findings-naacl.234","title":"Scaling Up Membership Inference: When and How Attacks Succeed on Large Language Models","display_name":"Scaling Up Membership Inference: When and How Attacks Succeed on Large Language Models","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4411120109","doi":"https://doi.org/10.18653/v1/2025.findings-naacl.234"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.findings-naacl.234","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-naacl.234","pdf_url":"https://aclanthology.org/2025.findings-naacl.234.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: NAACL 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-naacl.234.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5083791842","display_name":"Haritz Puerto","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Haritz Puerto","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5058150999","display_name":"Martin Gubri","orcid":"https://orcid.org/0000-0001-6744-6662"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Martin Gubri","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5003220288","display_name":"Sangdoo Yun","orcid":"https://orcid.org/0000-0002-0417-8450"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sangdoo Yun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5100605344","display_name":"Seung\u2010June Oh","orcid":"https://orcid.org/0000-0002-0322-3539"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Seong Joon Oh","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":11.9363,"has_fulltext":true,"cited_by_count":8,"citation_normalized_percentile":{"value":0.9869916,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":95,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"4165","last_page":"4182"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9474999904632568,"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/T10028","display_name":"Topic Modeling","score":0.9474999904632568,"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/computer-science","display_name":"Computer science","score":0.7210344076156616},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6966173648834229},{"id":"https://openalex.org/keywords/scaling","display_name":"Scaling","score":0.675565242767334},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.50438392162323},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.42812108993530273},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3916969895362854},{"id":"https://openalex.org/keywords/theoretical-computer-science","display_name":"Theoretical computer science","score":0.35149410367012024},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.13852781057357788}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7210344076156616},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6966173648834229},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.675565242767334},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.50438392162323},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.42812108993530273},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3916969895362854},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.35149410367012024},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.13852781057357788},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-naacl.234","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-naacl.234","pdf_url":"https://aclanthology.org/2025.findings-naacl.234.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: NAACL 2025","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-naacl.234","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-naacl.234","pdf_url":"https://aclanthology.org/2025.findings-naacl.234.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: NAACL 2025","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4411120109.pdf","grobid_xml":"https://content.openalex.org/works/W4411120109.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W141820298","https://openalex.org/W2049584446","https://openalex.org/W2079781215","https://openalex.org/W4378770497","https://openalex.org/W4308245303","https://openalex.org/W2014033564","https://openalex.org/W2055243143","https://openalex.org/W2910573937","https://openalex.org/W4385571583","https://openalex.org/W4389519396"],"abstract_inverted_index":{"Membership":[0],"inference":[1],"attacks":[2],"(MIA)":[3],"attempt":[4],"to":[5,72,126,166],"verify":[6],"the":[7,15,28,39,79,113,136,155,178],"membership":[8,159],"of":[9,31,41,78,120,129,133,138,157,173],"a":[10,19,117,127,147],"given":[11],"data":[12,121],"sample":[13],"in":[14,24],"training":[16,45],"set":[17],"for":[18,44,49,51,106,154],"model.MIA":[20],"has":[21,57],"become":[22],"relevant":[23],"recent":[25,55,148],"years,":[26],"following":[27],"rapid":[29],"development":[30],"large":[32],"language":[33],"models":[34],"(LLM).Many":[35],"are":[36,104],"concerned":[37],"about":[38],"usage":[40],"copyrighted":[42],"materials":[43],"them":[46],"and":[47,171,184],"call":[48],"methods":[50,63],"detecting":[52],"such":[53],"usage.However,":[54],"research":[56],"largely":[58],"concluded":[59],"that":[60,93,111,161],"current":[61,139],"MIA":[62,94,114,140,164,168,181],"do":[64],"not":[65],"work":[66,149],"on":[67,97,150,182],"LLMs.Even":[68],"when":[69,101],"they":[70],"seem":[71],"work,":[73,90],"it":[74],"is":[75],"usually":[76],"because":[77],"ill-designed":[80],"experimental":[81],"setup":[82],"where":[83],"other":[84],"shortcut":[85],"features":[86,165],"enable":[87,167],"\"cheating.\"In":[88],"this":[89],"we":[91,145],"argue":[92],"still":[95],"works":[96],"LLMs,":[98],"but":[99],"only":[100],"multiple":[102],"documents":[103,130,174],"presented":[105],"testing.We":[107],"construct":[108],"new":[109],"benchmarks":[110],"measure":[112],"performances":[115],"at":[116,142,169],"continuous":[118],"scale":[119],"samples,":[122],"from":[123],"sentences":[124],"(n-grams)":[125],"collection":[128,172],"(multiple":[131],"chunks":[132],"tokens).To":[134],"validate":[135],"efficacy":[137],"approaches":[141],"greater":[143],"scales,":[144],"adapt":[146],"Dataset":[151],"Inference":[152],"(DI)":[153],"task":[156],"binary":[158],"detection":[160],"aggregates":[162],"paragraph-level":[163],"document":[170],"level.This":[175],"baseline":[176],"achieves":[177],"first":[179],"successful":[180],"pre-trained":[183],"finetuned":[185],"LLMs.":[186],"1":[187]},"counts_by_year":[{"year":2026,"cited_by_count":6},{"year":2025,"cited_by_count":2}],"updated_date":"2026-08-11T07:18:39.950985","created_date":"2025-10-10T00:00:00"}