{"id":"https://openalex.org/W7106161175","doi":"https://doi.org/10.48550/arxiv.2511.14045","title":"Auditing Data Membership in Reinforcement Learning With Verifiable Rewards","display_name":"Auditing Data Membership in Reinforcement Learning With Verifiable Rewards","publication_year":2025,"publication_date":"2025-11-18","ids":{"openalex":"https://openalex.org/W7106161175","doi":"https://doi.org/10.48550/arxiv.2511.14045"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2511.14045","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2511.14045","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,"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2511.14045","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Liu, Yule","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Yule","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Zhang, Heyi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Heyi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Zheng, Jinyi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zheng, Jinyi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Sun, Zhen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sun, Zhen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Peng, Zifan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Peng, Zifan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Wei, Jiaheng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wei, Jiaheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Cong, Tianshuo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cong, Tianshuo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Yang, Yilong","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Yilong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":null,"display_name":"He, Xinlei","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"He, Xinlei","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":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.43720000982284546,"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":0.43720000982284546,"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":0.065700002014637,"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/T10883","display_name":"Ethics and Social Impacts of AI","score":0.05739999935030937,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/verifiable-secret-sharing","display_name":"Verifiable secret sharing","score":0.7063000202178955},{"id":"https://openalex.org/keywords/correctness","display_name":"Correctness","score":0.633400022983551},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6244999766349792},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.5156000256538391},{"id":"https://openalex.org/keywords/audit","display_name":"Audit","score":0.4293999969959259},{"id":"https://openalex.org/keywords/memorization","display_name":"Memorization","score":0.42309999465942383},{"id":"https://openalex.org/keywords/information-privacy","display_name":"Information privacy","score":0.3871999979019165},{"id":"https://openalex.org/keywords/audit-trail","display_name":"Audit trail","score":0.30230000615119934}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7307999730110168},{"id":"https://openalex.org/C85847156","wikidata":"https://www.wikidata.org/wiki/Q59015987","display_name":"Verifiable secret sharing","level":3,"score":0.7063000202178955},{"id":"https://openalex.org/C55439883","wikidata":"https://www.wikidata.org/wiki/Q360812","display_name":"Correctness","level":2,"score":0.633400022983551},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6244999766349792},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.5156000256538391},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4641999900341034},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4406999945640564},{"id":"https://openalex.org/C199521495","wikidata":"https://www.wikidata.org/wiki/Q181487","display_name":"Audit","level":2,"score":0.4293999969959259},{"id":"https://openalex.org/C30038468","wikidata":"https://www.wikidata.org/wiki/Q4354775","display_name":"Memorization","level":2,"score":0.42309999465942383},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.3995000123977661},{"id":"https://openalex.org/C123201435","wikidata":"https://www.wikidata.org/wiki/Q456632","display_name":"Information privacy","level":2,"score":0.3871999979019165},{"id":"https://openalex.org/C80958533","wikidata":"https://www.wikidata.org/wiki/Q1047174","display_name":"Audit trail","level":3,"score":0.30230000615119934},{"id":"https://openalex.org/C186108316","wikidata":"https://www.wikidata.org/wiki/Q352530","display_name":"Adaptive neuro fuzzy inference system","level":4,"score":0.29760000109672546},{"id":"https://openalex.org/C49937458","wikidata":"https://www.wikidata.org/wiki/Q2599292","display_name":"Probabilistic logic","level":2,"score":0.29170000553131104},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.2897000014781952},{"id":"https://openalex.org/C178489894","wikidata":"https://www.wikidata.org/wiki/Q8789","display_name":"Cryptography","level":2,"score":0.28369998931884766},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.28139999508857727},{"id":"https://openalex.org/C58166","wikidata":"https://www.wikidata.org/wiki/Q224821","display_name":"Fuzzy logic","level":2,"score":0.2741999924182892},{"id":"https://openalex.org/C111065885","wikidata":"https://www.wikidata.org/wiki/Q1189053","display_name":"Fuzz testing","level":3,"score":0.2736999988555908},{"id":"https://openalex.org/C46686674","wikidata":"https://www.wikidata.org/wiki/Q466303","display_name":"Boosting (machine learning)","level":2,"score":0.2597000002861023}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2511.14045","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2511.14045","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,"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":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2511.14045","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2511.14045","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,"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"},"sustainable_development_goals":[{"display_name":"Peace, Justice and strong institutions","id":"https://metadata.un.org/sdg/16","score":0.7745907306671143}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Reinforcement":[0],"Learning":[1],"with":[2,236],"Verifiable":[3],"Rewards":[4],"(RLVR)":[5],"has":[6],"become":[7],"a":[8,32,51,109,118,151,157,208],"core":[9],"training":[10,92,190],"stage":[11],"in":[12,132],"recent":[13],"large":[14],"language":[15],"models":[16],"(LLMs).":[17],"Its":[18],"reliance":[19],"on":[20,91,204],"non-public,":[21],"high-value":[22],"prompt":[23],"sets":[24],"raises":[25],"concerns":[26],"about":[27],"unauthorized":[28],"data":[29,77],"use,":[30],"creating":[31],"need":[33],"for":[34,114],"exposure":[35],"auditing.":[36,103],"A":[37],"natural":[38],"tool":[39],"is":[40,187,213],"membership":[41],"inference":[42],"attacks":[43],"(MIAs),":[44],"but":[45,224],"existing":[46],"methods":[47],"detect":[48],"fitting":[49],"to":[50,59],"fixed":[52],"target":[53],"string.":[54],"This":[55],"does":[56],"not":[57],"apply":[58],"RLVR,":[60],"which":[61],"generates":[62],"responses":[63],"from":[64],"the":[65,74,87,198,205,220],"model":[66,120,200,217],"itself":[67],"and":[68,136,169,176,195,229],"reinforces":[69],"successful":[70],"ones,":[71],"thus":[72],"hindering":[73],"auditing":[75,112,154,186],"of":[76],"exposure.":[78],"We":[79,104,181],"show":[80,183],"that":[81,97,184],"it":[82],"remains":[83],"detectable:":[84],"RLVR":[85,185,222],"reshapes":[86],"model's":[88],"response":[89],"distribution":[90,234],"prompts,":[93],"producing":[94],"behavioral":[95,141],"traces":[96,194],"can":[98,230],"be":[99],"surfaced":[100],"through":[101],"targeted":[102],"propose":[105],"Divergence-in-Behavior":[106],"Auditing":[107],"(DIBA),":[108],"white-box":[110,158],"query-level":[111,153],"framework":[113],"RLVR.":[115],"DIBA":[116,149,160],"compares":[117],"fine-tuned":[119],"against":[121],"its":[122],"pre-RLVR":[123],"checkpoint":[124],"along":[125],"two":[126],"axes:":[127],"reward-side":[128],"evidence":[129,138],"capturing":[130,139],"changes":[131],"verifiable":[133],"task":[134],"success,":[135],"policy-side":[137],"prompt-conditioned":[140],"drift.":[142],"By":[143],"aggregating":[144],"over":[145],"multiple":[146],"stochastic":[147],"rollouts,":[148],"produces":[150],"stable":[152],"signal.":[155],"Under":[156,207],"setting,":[159,211],"consistently":[161],"outperforms":[162],"strong":[163],"transferred":[164],"likelihood-based":[165],"baselines,":[166],"including":[167],"calibrated":[168],"self-generated":[170],"variants,":[171],"achieving":[172],"around":[173],"0.8":[174],"AUC":[175],"an":[177],"order-of-magnitude":[178],"stronger":[179,188],"TPR@0.1%FPR.":[180],"further":[182],"when":[189,197],"leaves":[191],"non-trivial":[192],"prompt-specific":[193],"weaker":[196],"base":[199],"already":[201],"performs":[202],"well":[203],"prompt.":[206],"practical":[209],"grey-box":[210],"transfer":[212],"often":[214],"robust":[215],"across":[216,227],"sizes":[218],"under":[219,233],"same":[221],"algorithm,":[223],"more":[225],"varied":[226],"algorithms,":[228],"remain":[231],"useful":[232],"shift":[235],"carefully":[237],"chosen":[238],"shadow":[239],"data.":[240]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2025-11-20T00:00:00"}