{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:VHEX3QEUQB4XPINB7JEOP2SYLO","short_pith_number":"pith:VHEX3QEU","schema_version":"1.0","canonical_sha256":"a9c97dc094807977a1a1fa48e7ea585bbd0f9886fb9dddf5b257f84a055327e6","source":{"kind":"arxiv","id":"2405.20446","version":3},"attestation_state":"computed","paper":{"title":"Is My Data in Your Retrieval Database? Membership Inference Attacks Against Retrieval Augmented Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Abigail Goldsteen, Guy Amit, Maya Anderson","submitted_at":"2024-05-30T19:46:36Z","abstract_excerpt":"Retrieval Augmented Generation (RAG) systems have shown great promise in natural language processing. However, their reliance on data stored in a retrieval database, which may contain proprietary or sensitive information, introduces new privacy concerns. Specifically, an attacker may be able to infer whether a certain text passage appears in the retrieval database by observing the outputs of the RAG system, an attack known as a Membership Inference Attack (MIA). Despite the significance of this threat, MIAs against RAG systems have yet remained under-explored. This study addresses this gap by "},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2405.20446","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-05-30T19:46:36Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"3ace5465540fd6cda7d3e8a3a369a431a07cec0c7d07b9b44936a2b015e3a3e0","abstract_canon_sha256":"4d1f42d1076766ab6e0f7bf6b76b08138a9a9a6b29fe75448236cb5d1006eef6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:26:03.980620Z","signature_b64":"GCIJqNwRhjnFQ8n13syCgmxBc4ZX8aDLWBDJOv67xSeBoCQAnx5yMnt+krzVmGPl+eI6VDsEpRuoIUEMP0NMBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a9c97dc094807977a1a1fa48e7ea585bbd0f9886fb9dddf5b257f84a055327e6","last_reissued_at":"2026-07-05T10:26:03.980102Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:26:03.980102Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Is My Data in Your Retrieval Database? Membership Inference Attacks Against Retrieval Augmented Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Abigail Goldsteen, Guy Amit, Maya Anderson","submitted_at":"2024-05-30T19:46:36Z","abstract_excerpt":"Retrieval Augmented Generation (RAG) systems have shown great promise in natural language processing. However, their reliance on data stored in a retrieval database, which may contain proprietary or sensitive information, introduces new privacy concerns. Specifically, an attacker may be able to infer whether a certain text passage appears in the retrieval database by observing the outputs of the RAG system, an attack known as a Membership Inference Attack (MIA). Despite the significance of this threat, MIAs against RAG systems have yet remained under-explored. This study addresses this gap by "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.20446","kind":"arxiv","version":3},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2405.20446/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2405.20446","created_at":"2026-07-05T10:26:03.980163+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.20446v3","created_at":"2026-07-05T10:26:03.980163+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.20446","created_at":"2026-07-05T10:26:03.980163+00:00"},{"alias_kind":"pith_short_12","alias_value":"VHEX3QEUQB4X","created_at":"2026-07-05T10:26:03.980163+00:00"},{"alias_kind":"pith_short_16","alias_value":"VHEX3QEUQB4XPINB","created_at":"2026-07-05T10:26:03.980163+00:00"},{"alias_kind":"pith_short_8","alias_value":"VHEX3QEU","created_at":"2026-07-05T10:26:03.980163+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26627","citing_title":"Agents That Know Too Much: A Data-Centric Survey of Privacy in LLM Agents","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2606.05787","citing_title":"SentinelRAG: Synthetic Sentinel Knowledge for RAG Database Copyright Protection","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2409.10102","citing_title":"Trustworthiness in Retrieval-Augmented Generation Systems: A Survey","ref_index":86,"is_internal_anchor":false},{"citing_arxiv_id":"2510.18333","citing_title":"Position: LLM Watermarking Should Align Stakeholders' Incentives for Practical Adoption","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2603.09002","citing_title":"Security Considerations for Multi-agent Systems","ref_index":259,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VHEX3QEUQB4XPINB7JEOP2SYLO","json":"https://pith.science/pith/VHEX3QEUQB4XPINB7JEOP2SYLO.json","graph_json":"https://pith.science/api/pith-number/VHEX3QEUQB4XPINB7JEOP2SYLO/graph.json","events_json":"https://pith.science/api/pith-number/VHEX3QEUQB4XPINB7JEOP2SYLO/events.json","paper":"https://pith.science/paper/VHEX3QEU"},"agent_actions":{"view_html":"https://pith.science/pith/VHEX3QEUQB4XPINB7JEOP2SYLO","download_json":"https://pith.science/pith/VHEX3QEUQB4XPINB7JEOP2SYLO.json","view_paper":"https://pith.science/paper/VHEX3QEU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.20446&json=true","fetch_graph":"https://pith.science/api/pith-number/VHEX3QEUQB4XPINB7JEOP2SYLO/graph.json","fetch_events":"https://pith.science/api/pith-number/VHEX3QEUQB4XPINB7JEOP2SYLO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VHEX3QEUQB4XPINB7JEOP2SYLO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VHEX3QEUQB4XPINB7JEOP2SYLO/action/storage_attestation","attest_author":"https://pith.science/pith/VHEX3QEUQB4XPINB7JEOP2SYLO/action/author_attestation","sign_citation":"https://pith.science/pith/VHEX3QEUQB4XPINB7JEOP2SYLO/action/citation_signature","submit_replication":"https://pith.science/pith/VHEX3QEUQB4XPINB7JEOP2SYLO/action/replication_record"}},"created_at":"2026-07-05T10:26:03.980163+00:00","updated_at":"2026-07-05T10:26:03.980163+00:00"}