{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:UA3XGK77LUFHRMGELCJL3TGS4K","short_pith_number":"pith:UA3XGK77","canonical_record":{"source":{"id":"2402.17012","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-02-26T20:41:50Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"150425b500b5e9e32550ad471f64d1cf55fd4b940e85b99a4096c963d046e864","abstract_canon_sha256":"b06019ce5b75992b2a951189cca889496b638d2340b8e0b9215822312bf60218"},"schema_version":"1.0"},"canonical_sha256":"a037732bff5d0a78b0c45892bdccd2e28cc2299e08d045df599b51d75ddc10ba","source":{"kind":"arxiv","id":"2402.17012","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.17012","created_at":"2026-07-05T08:43:53Z"},{"alias_kind":"arxiv_version","alias_value":"2402.17012v4","created_at":"2026-07-05T08:43:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.17012","created_at":"2026-07-05T08:43:53Z"},{"alias_kind":"pith_short_12","alias_value":"UA3XGK77LUFH","created_at":"2026-07-05T08:43:53Z"},{"alias_kind":"pith_short_16","alias_value":"UA3XGK77LUFHRMGE","created_at":"2026-07-05T08:43:53Z"},{"alias_kind":"pith_short_8","alias_value":"UA3XGK77","created_at":"2026-07-05T08:43:53Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:UA3XGK77LUFHRMGELCJL3TGS4K","target":"record","payload":{"canonical_record":{"source":{"id":"2402.17012","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-02-26T20:41:50Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"150425b500b5e9e32550ad471f64d1cf55fd4b940e85b99a4096c963d046e864","abstract_canon_sha256":"b06019ce5b75992b2a951189cca889496b638d2340b8e0b9215822312bf60218"},"schema_version":"1.0"},"canonical_sha256":"a037732bff5d0a78b0c45892bdccd2e28cc2299e08d045df599b51d75ddc10ba","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:43:53.496279Z","signature_b64":"/fCR5EiIITwOQSbgtwlcs8WXDh1GvyxNMrfO9M01jfxjQraK94cEDH6d/v/nTOkvRkzWUOKa8OWGSrjebFTnAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a037732bff5d0a78b0c45892bdccd2e28cc2299e08d045df599b51d75ddc10ba","last_reissued_at":"2026-07-05T08:43:53.495805Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:43:53.495805Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2402.17012","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:43:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Uv3rvht9429QYFJREtu7/k1b+W1T9n1i9zTvDbi61ZCR8YCnVDGDg4cymYAwhzdbxgpr1OPyRcWRMM5PAMzrDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T13:06:38.556402Z"},"content_sha256":"985989e85a79180fed600a708362d8d6ca8419cad8833c66e248138171b126dc","schema_version":"1.0","event_id":"sha256:985989e85a79180fed600a708362d8d6ca8419cad8833c66e248138171b126dc"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:UA3XGK77LUFHRMGELCJL3TGS4K","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Pandora's White-Box: Precise Training Data Detection and Extraction in Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Jason Wang, Jeffrey G. Wang, Marvin Li, Seth Neel","submitted_at":"2024-02-26T20:41:50Z","abstract_excerpt":"In this paper we develop state-of-the-art privacy attacks against Large Language Models (LLMs), where an adversary with some access to the model tries to learn something about the underlying training data. Our headline results are new membership inference attacks (MIAs) against pretrained LLMs that perform hundreds of times better than baseline attacks, and a pipeline showing that over 50% (!) of the fine-tuning dataset can be extracted from a fine-tuned LLM in natural settings. We consider varying degrees of access to the underlying model, pretraining and fine-tuning data, and both MIAs and t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.17012","kind":"arxiv","version":4},"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/2402.17012/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:43:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rCwUTy9i/aM6oPdYlIV0i7m9cuaF9u5+WzNxGYCYzZ/TW/9CXTZJu2W37MJflGYjn84vh7RA0Px1j1FgOja4Dw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T13:06:38.556921Z"},"content_sha256":"31aec203b69a5095681caf13a0b4e836548f67cde5cf1638ff070c6edd66bb1f","schema_version":"1.0","event_id":"sha256:31aec203b69a5095681caf13a0b4e836548f67cde5cf1638ff070c6edd66bb1f"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/UA3XGK77LUFHRMGELCJL3TGS4K/bundle.json","state_url":"https://pith.science/pith/UA3XGK77LUFHRMGELCJL3TGS4K/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/UA3XGK77LUFHRMGELCJL3TGS4K/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-10T13:06:38Z","links":{"resolver":"https://pith.science/pith/UA3XGK77LUFHRMGELCJL3TGS4K","bundle":"https://pith.science/pith/UA3XGK77LUFHRMGELCJL3TGS4K/bundle.json","state":"https://pith.science/pith/UA3XGK77LUFHRMGELCJL3TGS4K/state.json","well_known_bundle":"https://pith.science/.well-known/pith/UA3XGK77LUFHRMGELCJL3TGS4K/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:UA3XGK77LUFHRMGELCJL3TGS4K","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"b06019ce5b75992b2a951189cca889496b638d2340b8e0b9215822312bf60218","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-02-26T20:41:50Z","title_canon_sha256":"150425b500b5e9e32550ad471f64d1cf55fd4b940e85b99a4096c963d046e864"},"schema_version":"1.0","source":{"id":"2402.17012","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.17012","created_at":"2026-07-05T08:43:53Z"},{"alias_kind":"arxiv_version","alias_value":"2402.17012v4","created_at":"2026-07-05T08:43:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.17012","created_at":"2026-07-05T08:43:53Z"},{"alias_kind":"pith_short_12","alias_value":"UA3XGK77LUFH","created_at":"2026-07-05T08:43:53Z"},{"alias_kind":"pith_short_16","alias_value":"UA3XGK77LUFHRMGE","created_at":"2026-07-05T08:43:53Z"},{"alias_kind":"pith_short_8","alias_value":"UA3XGK77","created_at":"2026-07-05T08:43:53Z"}],"graph_snapshots":[{"event_id":"sha256:31aec203b69a5095681caf13a0b4e836548f67cde5cf1638ff070c6edd66bb1f","target":"graph","created_at":"2026-07-05T08:43:53Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2402.17012/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"In this paper we develop state-of-the-art privacy attacks against Large Language Models (LLMs), where an adversary with some access to the model tries to learn something about the underlying training data. Our headline results are new membership inference attacks (MIAs) against pretrained LLMs that perform hundreds of times better than baseline attacks, and a pipeline showing that over 50% (!) of the fine-tuning dataset can be extracted from a fine-tuned LLM in natural settings. We consider varying degrees of access to the underlying model, pretraining and fine-tuning data, and both MIAs and t","authors_text":"Jason Wang, Jeffrey G. Wang, Marvin Li, Seth Neel","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-02-26T20:41:50Z","title":"Pandora's White-Box: Precise Training Data Detection and Extraction in Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.17012","kind":"arxiv","version":4},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:985989e85a79180fed600a708362d8d6ca8419cad8833c66e248138171b126dc","target":"record","created_at":"2026-07-05T08:43:53Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"b06019ce5b75992b2a951189cca889496b638d2340b8e0b9215822312bf60218","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-02-26T20:41:50Z","title_canon_sha256":"150425b500b5e9e32550ad471f64d1cf55fd4b940e85b99a4096c963d046e864"},"schema_version":"1.0","source":{"id":"2402.17012","kind":"arxiv","version":4}},"canonical_sha256":"a037732bff5d0a78b0c45892bdccd2e28cc2299e08d045df599b51d75ddc10ba","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a037732bff5d0a78b0c45892bdccd2e28cc2299e08d045df599b51d75ddc10ba","first_computed_at":"2026-07-05T08:43:53.495805Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:43:53.495805Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"/fCR5EiIITwOQSbgtwlcs8WXDh1GvyxNMrfO9M01jfxjQraK94cEDH6d/v/nTOkvRkzWUOKa8OWGSrjebFTnAg==","signature_status":"signed_v1","signed_at":"2026-07-05T08:43:53.496279Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.17012","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:985989e85a79180fed600a708362d8d6ca8419cad8833c66e248138171b126dc","sha256:31aec203b69a5095681caf13a0b4e836548f67cde5cf1638ff070c6edd66bb1f"],"state_sha256":"cf54f8dce71083b25539dd77e9e9a479f9357ad109362096117aa87b45f4a33c"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"p3EuOduiUA7tRT8iNeGGUZXy1pgE96jrJceiVTQL7TPLGkb57a3l7Jmv1DOMj5Grc9ROSa9RhPd/CxqQ4qFDDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-10T13:06:38.560780Z","bundle_sha256":"05bd0af1f39afad87a80b4e947f0122af2e01a1d179f5fdf23cfe4792ff4d8aa"}}