{"id":"https://openalex.org/W4387985331","doi":"https://doi.org/10.48550/arxiv.2310.16789","title":"Detecting Pretraining Data from Large Language Models","display_name":"Detecting Pretraining Data from Large Language Models","alternate_titles":[],"publication_year":2023,"publication_date":"2023-10-25","ids":{"openalex":"https://openalex.org/W4387985331","doi":"https://doi.org/10.48550/arxiv.2310.16789"},"language":"en","primary_location":{"id":"pmh:oai:proceedings.iclr.cc:2024/e32ad85fa27be4a9868d55703f01323e","doi":null,"is_oa":true,"landing_page_url":"https://proceedings.iclr.cc/paper_files/paper/2024/hash/e32ad85fa27be4a9868d55703f01323e-Abstract-Conference.html","pdf_url":null,"source":{"id":"https://openalex.org/S4306419637","display_name":"International Conference on Learning Representations","issn_l":null,"issn":null,"is_oa":true,"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":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"International Conference on Learning Representations 2024","raw_type":"info:eu-repo/semantics/conferenceObject"},"parent_works":[],"lineage":["https://openalex.org/W4387985331"],"child_works_count":0,"type":"conference-paper","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://proceedings.iclr.cc/paper_files/paper/2024/hash/e32ad85fa27be4a9868d55703f01323e-Abstract-Conference.html","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5103238719","display_name":"Weijia Shi","orcid":"https://orcid.org/0000-0002-3379-0331","observed_orcids":["https://orcid.org/0000-0002-3379-0331"]},"institutions":[{"id":"https://openalex.org/I201448701","display_name":"University of Washington","ror":"https://ror.org/00cvxb145","country_code":"US","type":"education","lineage":["https://openalex.org/I201448701"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Shi, Weijia","raw_affiliation_strings":["University of Washington"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Washington","institution_ids":["https://openalex.org/I201448701"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5000037569","display_name":"Anirudh Ajith","orcid":"https://orcid.org/0009-0003-0447-588X","observed_orcids":["https://orcid.org/0009-0003-0447-588X"]},"institutions":[{"id":"https://openalex.org/I20089843","display_name":"Princeton University","ror":"https://ror.org/00hx57361","country_code":"US","type":"education","lineage":["https://openalex.org/I20089843"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Ajith, Anirudh","raw_affiliation_strings":["Princeton University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Princeton University","institution_ids":["https://openalex.org/I20089843"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5003465314","display_name":"Mengzhou Xia","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I20089843","display_name":"Princeton University","ror":"https://ror.org/00hx57361","country_code":"US","type":"education","lineage":["https://openalex.org/I20089843"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Xia, Mengzhou","raw_affiliation_strings":["Princeton University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Princeton University","institution_ids":["https://openalex.org/I20089843"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5082426637","display_name":"Yangsibo Huang","orcid":"https://orcid.org/0000-0002-0640-4845","observed_orcids":["https://orcid.org/0000-0002-0640-4845"]},"institutions":[{"id":"https://openalex.org/I20089843","display_name":"Princeton University","ror":"https://ror.org/00hx57361","country_code":"US","type":"education","lineage":["https://openalex.org/I20089843"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Huang, Yangsibo","raw_affiliation_strings":["Princeton University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Princeton University","institution_ids":["https://openalex.org/I20089843"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5005358671","display_name":"Daogao Liu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I201448701","display_name":"University of Washington","ror":"https://ror.org/00cvxb145","country_code":"US","type":"education","lineage":["https://openalex.org/I201448701"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Liu, Daogao","raw_affiliation_strings":["University of Washington"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Washington","institution_ids":["https://openalex.org/I201448701"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5088311840","display_name":"Terra Blevins","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I201448701","display_name":"University of Washington","ror":"https://ror.org/00cvxb145","country_code":"US","type":"education","lineage":["https://openalex.org/I201448701"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Blevins, Terra","raw_affiliation_strings":["University of Washington"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Washington","institution_ids":["https://openalex.org/I201448701"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5051064208","display_name":"Danqi Chen","orcid":"https://orcid.org/0000-0002-6226-6838","observed_orcids":["https://orcid.org/0000-0002-6226-6838"]},"institutions":[{"id":"https://openalex.org/I20089843","display_name":"Princeton University","ror":"https://ror.org/00hx57361","country_code":"US","type":"education","lineage":["https://openalex.org/I20089843"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Chen, Danqi","raw_affiliation_strings":["Princeton University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Princeton University","institution_ids":["https://openalex.org/I20089843"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5067919401","display_name":"Luke Zettlemoyer","orcid":"https://orcid.org/0009-0008-8296-0764","observed_orcids":["https://orcid.org/0009-0008-8296-0764"]},"institutions":[{"id":"https://openalex.org/I201448701","display_name":"University of Washington","ror":"https://ror.org/00cvxb145","country_code":"US","type":"education","lineage":["https://openalex.org/I201448701"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Zettlemoyer, Luke","raw_affiliation_strings":["University of Washington"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Washington","institution_ids":["https://openalex.org/I201448701"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":29.4534,"has_fulltext":true,"cited_by_count":505,"citation_normalized_percentile":{"value":0.9970826,"is_in_top_1_percent":true,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":98,"max":100},"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_pulled":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.3434999883174896,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.3434999883174896,"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":0.2946999967098236,"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/T12380","display_name":"Authorship Attribution and Profiling","score":0.10520000010728836,"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-models","display_name":"large language models","score":0.9769999980926514},{"id":"https://openalex.org/keywords/black-box-detection","display_name":"black-box detection","score":0.5870000123977661},{"id":"https://openalex.org/keywords/privacy-auditing","display_name":"privacy auditing","score":0.6389999985694885},{"id":"https://openalex.org/keywords/machine-unlearning","display_name":"machine unlearning","score":0.9869999885559082}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7439150214195251},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.7104443311691284},{"id":"https://openalex.org/C2776145971","wikidata":"https://www.wikidata.org/wiki/Q30673951","display_name":"Labeled data","level":2,"score":0.5622666478157043},{"id":"https://openalex.org/C79337645","wikidata":"https://www.wikidata.org/wiki/Q779824","display_name":"Outlier","level":2,"score":0.5592784881591797},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.455265074968338},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4531933069229126},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.45302814245224},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.45086437463760376},{"id":"https://openalex.org/C739882","wikidata":"https://www.wikidata.org/wiki/Q3560506","display_name":"Anomaly detection","level":2,"score":0.4443434178829193},{"id":"https://openalex.org/C2777267654","wikidata":"https://www.wikidata.org/wiki/Q3519023","display_name":"Test (biology)","level":2,"score":0.4140355885028839},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.40218162536621094},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.32170283794403076},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0},{"id":"https://openalex.org/C151730666","wikidata":"https://www.wikidata.org/wiki/Q7205","display_name":"Paleontology","level":1,"score":0.0},{"id":"https://openalex.org/C111472728","wikidata":"https://www.wikidata.org/wiki/Q9471","display_name":"Epistemology","level":1,"score":0.0},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.0},{"id":"https://openalex.org/C205649164","wikidata":"https://www.wikidata.org/wiki/Q1071","display_name":"Geography","level":0,"score":0.0},{"id":"https://openalex.org/C13280743","wikidata":"https://www.wikidata.org/wiki/Q131089","display_name":"Geodesy","level":1,"score":0.0}],"mesh":[],"locations_count":3,"locations":[{"id":"pmh:oai:proceedings.iclr.cc:2024/e32ad85fa27be4a9868d55703f01323e","doi":null,"is_oa":true,"landing_page_url":"https://proceedings.iclr.cc/paper_files/paper/2024/hash/e32ad85fa27be4a9868d55703f01323e-Abstract-Conference.html","pdf_url":null,"source":{"id":"https://openalex.org/S4306419637","display_name":"International Conference on Learning Representations","issn_l":null,"issn":null,"is_oa":true,"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":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"International Conference on Learning Representations 2024","raw_type":"info:eu-repo/semantics/conferenceObject"},{"id":"pmh:oai:arXiv.org:2310.16789","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2310.16789","pdf_url":"https://arxiv.org/pdf/2310.16789","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,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2310.16789","doi":"https://doi.org/10.48550/arxiv.2310.16789","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2310.16789","pdf_url":null,"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,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:proceedings.iclr.cc:2024/e32ad85fa27be4a9868d55703f01323e","doi":null,"is_oa":true,"landing_page_url":"https://proceedings.iclr.cc/paper_files/paper/2024/hash/e32ad85fa27be4a9868d55703f01323e-Abstract-Conference.html","pdf_url":null,"source":{"id":"https://openalex.org/S4306419637","display_name":"International Conference on Learning Representations","issn_l":null,"issn":null,"is_oa":true,"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":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"International Conference on Learning Representations 2024","raw_type":"info:eu-repo/semantics/conferenceObject"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4387985331.pdf","grobid_xml":"https://content.openalex.org/works/W4387985331.grobid-xml"},"referenced_works_count":49,"referenced_works":["https://openalex.org/W2113459411","https://openalex.org/W2154652894","https://openalex.org/W2535690855","https://openalex.org/W2786233556","https://openalex.org/W2795435272","https://openalex.org/W2947693768","https://openalex.org/W2950181023","https://openalex.org/W2952604841","https://openalex.org/W2961073278","https://openalex.org/W2963794891","https://openalex.org/W2963995027","https://openalex.org/W3027379683","https://openalex.org/W3030163527","https://openalex.org/W3034452044","https://openalex.org/W3048775464","https://openalex.org/W3101678025","https://openalex.org/W3112689365","https://openalex.org/W3116939721","https://openalex.org/W3118781290","https://openalex.org/W3120806613","https://openalex.org/W3135378441","https://openalex.org/W3153896080","https://openalex.org/W3154155772","https://openalex.org/W3156636935","https://openalex.org/W3173769540","https://openalex.org/W4224051134","https://openalex.org/W4224308101","https://openalex.org/W4226142937","https://openalex.org/W4229005866","https://openalex.org/W4288057780","https://openalex.org/W4309163068","https://openalex.org/W4312874698","https://openalex.org/W4318351452","https://openalex.org/W4321276312","https://openalex.org/W4322718191","https://openalex.org/W4327810158","https://openalex.org/W4362655426","https://openalex.org/W4385565597","https://openalex.org/W4385573947","https://openalex.org/W4385714201","https://openalex.org/W4385965989","https://openalex.org/W4386185470","https://openalex.org/W4387389584","https://openalex.org/W4389519044","https://openalex.org/W4391382837","https://openalex.org/W7133200591","https://openalex.org/W7133219637","https://openalex.org/W7133237827","https://openalex.org/W7133260402"],"related_works":["https://openalex.org/W6960412607","https://openalex.org/W4402671929","https://openalex.org/W4402670804","https://openalex.org/W4404784117","https://openalex.org/W4416044422","https://openalex.org/W4401306462","https://openalex.org/W4414360689","https://openalex.org/W4416035129","https://openalex.org/W7164025055","https://openalex.org/W7134060926"],"abstract_inverted_index":{"Although":[0],"large":[1],"language":[2],"models":[3],"(LLMs)":[4],"are":[5],"widely":[6,51],"deployed,":[7],"the":[8,18,79,97,104,109,165,190,214],"data":[9,49,64,81,124,209],"used":[10],"to":[11,25,61,92,131,155,174,213,238],"train":[12],"them":[13],"is":[14,30,68,153,171,211],"rarely":[15],"disclosed.":[16],"Given":[17],"incredible":[19],"scale":[20],"of":[21,27,65,87,252],"this":[22,75,114],"data,":[23,99],"up":[24],"trillions":[26],"tokens,":[28],"it":[29,35,257],"all":[31],"but":[32],"certain":[33],"that":[34,122,202,210,221],"includes":[36],"potentially":[37],"problematic":[38],"text":[39,88],"such":[40,178],"as":[41],"copyrighted":[42,242],"materials,":[43],"personally":[44],"identifiable":[45],"information,":[46],"and":[47,89,127,249,255],"test":[48],"for":[50],"reported":[52],"reference":[53,206],"benchmarks.":[54],"However,":[55],"we":[56,77,101,116],"currently":[57],"have":[58,175],"no":[59],"way":[60],"know":[62],"which":[63],"these":[66,231],"types":[67],"included":[69],"or":[70,193],"in":[71],"what":[72],"proportions.":[73],"In":[74],"paper,":[76],"study":[78],"pretraining":[80,98,191,215],"detection":[82,141,200,248],"problem:":[83],"given":[84],"a":[85,118,139,147,157,168,205,225,258],"piece":[86],"black-box":[90],"access":[91],"an":[93,150],"LLM":[94],"without":[95,186],"knowing":[96],"can":[100,183],"determine":[102],"if":[103],"model":[105,129,207],"was":[106],"trained":[107],"on":[108,146,208,228],"provided":[110],"text?":[111],"To":[112],"facilitate":[113],"study,":[115],"introduce":[117,138],"dynamic":[119],"benchmark":[120],"WIKIMIA":[121,229],"uses":[123],"created":[125],"before":[126],"after":[128],"training":[130,204],"support":[132],"gold":[133],"truth":[134],"detection.":[135],"We":[136,234],"also":[137],"new":[140],"method":[142],"Min-K%":[143,181,222,236],"Prob":[144,182,223,237],"based":[145],"simple":[148],"hypothesis:":[149],"unseen":[151],"example":[152,170,247],"likely":[154,173],"contain":[156],"few":[158],"outlier":[159],"words":[160,176],"with":[161,177],"low":[162,179],"probabilities":[163],"under":[164],"LLM,":[166],"while":[167],"seen":[169],"less":[172],"probabilities.":[180],"be":[184],"applied":[185],"any":[187,194],"knowledge":[188],"about":[189],"corpus":[192],"additional":[195],"training,":[196],"departing":[197],"from":[198],"previous":[199,232],"methods":[201],"require":[203],"similar":[212],"data.":[216],"Moreover,":[217],"our":[218],"experiments":[219],"demonstrate":[220],"achieves":[224],"7.4%":[226],"improvement":[227],"over":[230],"methods.":[233],"apply":[235],"three":[239],"real-world":[240],"scenarios,":[241],"book":[243],"detection,":[244],"contaminated":[245],"downstream":[246],"privacy":[250],"auditing":[251],"machine":[253],"unlearning,":[254],"find":[256],"consistently":[259],"effective":[260],"solution.":[261]},"cited_by_api_url":"https://api.openalex.org/works?filter=cites:W4387985331","counts_by_year":[{"year":2026,"cited_by_count":170},{"year":2025,"cited_by_count":195},{"year":2024,"cited_by_count":133},{"year":2023,"cited_by_count":7}],"updated_date":"2026-11-20T13:11:24.684918","created_date":"2023-10-28T00:00:00"}
