{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:IVPBXOGUMPPBOX3R6PMH4NISBG","short_pith_number":"pith:IVPBXOGU","schema_version":"1.0","canonical_sha256":"455e1bb8d463de175f71f3d87e351209af3c2d0c23307a9a5294e52a1c14025d","source":{"kind":"arxiv","id":"2409.14513","version":2},"attestation_state":"computed","paper":{"title":"Order of Magnitude Speedups for LLM Membership Inference","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR","stat.ML"],"primary_cat":"cs.LG","authors_text":"Aaron Roth, Martin Bertran, Rongting Zhang","submitted_at":"2024-09-22T16:18:14Z","abstract_excerpt":"Large Language Models (LLMs) have the promise to revolutionize computing broadly, but their complexity and extensive training data also expose significant privacy vulnerabilities. One of the simplest privacy risks associated with LLMs is their susceptibility to membership inference attacks (MIAs), wherein an adversary aims to determine whether a specific data point was part of the model's training set. Although this is a known risk, state of the art methodologies for MIAs rely on training multiple computationally costly shadow models, making risk evaluation prohibitive for large models. Here w"},"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":"2409.14513","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-09-22T16:18:14Z","cross_cats_sorted":["cs.CR","stat.ML"],"title_canon_sha256":"5ad2ce011491948408c1061d68e8173f2fb53d48447f2b153ccda221f7cd2b26","abstract_canon_sha256":"6dfb0c4067d56b5e7219c6879ac4183ac4e0d229f7898c7d2e5bef695d251b3e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:11:17.000247Z","signature_b64":"4Y/krtAQJ8Q0lAv8B2Yfph2qb+m5A8UGl1KCWuvUPCh/EJd9J1gFWWswDxR66qftXdtdCq0fCOM5QUfKiU2tDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"455e1bb8d463de175f71f3d87e351209af3c2d0c23307a9a5294e52a1c14025d","last_reissued_at":"2026-07-05T09:11:16.999806Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:11:16.999806Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Order of Magnitude Speedups for LLM Membership Inference","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR","stat.ML"],"primary_cat":"cs.LG","authors_text":"Aaron Roth, Martin Bertran, Rongting Zhang","submitted_at":"2024-09-22T16:18:14Z","abstract_excerpt":"Large Language Models (LLMs) have the promise to revolutionize computing broadly, but their complexity and extensive training data also expose significant privacy vulnerabilities. One of the simplest privacy risks associated with LLMs is their susceptibility to membership inference attacks (MIAs), wherein an adversary aims to determine whether a specific data point was part of the model's training set. Although this is a known risk, state of the art methodologies for MIAs rely on training multiple computationally costly shadow models, making risk evaluation prohibitive for large models. Here w"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.14513","kind":"arxiv","version":2},"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/2409.14513/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":"2409.14513","created_at":"2026-07-05T09:11:16.999864+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.14513v2","created_at":"2026-07-05T09:11:16.999864+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.14513","created_at":"2026-07-05T09:11:16.999864+00:00"},{"alias_kind":"pith_short_12","alias_value":"IVPBXOGUMPPB","created_at":"2026-07-05T09:11:16.999864+00:00"},{"alias_kind":"pith_short_16","alias_value":"IVPBXOGUMPPBOX3R","created_at":"2026-07-05T09:11:16.999864+00:00"},{"alias_kind":"pith_short_8","alias_value":"IVPBXOGU","created_at":"2026-07-05T09:11:16.999864+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.26133","citing_title":"Pretraining Data Exposure in Large Language Models: A Survey of Membership Inference, Data Contamination, and Security Implications","ref_index":63,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IVPBXOGUMPPBOX3R6PMH4NISBG","json":"https://pith.science/pith/IVPBXOGUMPPBOX3R6PMH4NISBG.json","graph_json":"https://pith.science/api/pith-number/IVPBXOGUMPPBOX3R6PMH4NISBG/graph.json","events_json":"https://pith.science/api/pith-number/IVPBXOGUMPPBOX3R6PMH4NISBG/events.json","paper":"https://pith.science/paper/IVPBXOGU"},"agent_actions":{"view_html":"https://pith.science/pith/IVPBXOGUMPPBOX3R6PMH4NISBG","download_json":"https://pith.science/pith/IVPBXOGUMPPBOX3R6PMH4NISBG.json","view_paper":"https://pith.science/paper/IVPBXOGU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.14513&json=true","fetch_graph":"https://pith.science/api/pith-number/IVPBXOGUMPPBOX3R6PMH4NISBG/graph.json","fetch_events":"https://pith.science/api/pith-number/IVPBXOGUMPPBOX3R6PMH4NISBG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IVPBXOGUMPPBOX3R6PMH4NISBG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IVPBXOGUMPPBOX3R6PMH4NISBG/action/storage_attestation","attest_author":"https://pith.science/pith/IVPBXOGUMPPBOX3R6PMH4NISBG/action/author_attestation","sign_citation":"https://pith.science/pith/IVPBXOGUMPPBOX3R6PMH4NISBG/action/citation_signature","submit_replication":"https://pith.science/pith/IVPBXOGUMPPBOX3R6PMH4NISBG/action/replication_record"}},"created_at":"2026-07-05T09:11:16.999864+00:00","updated_at":"2026-07-05T09:11:16.999864+00:00"}