{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:ASSWOMG3RWDG4VY4YVLT4H7VCV","short_pith_number":"pith:ASSWOMG3","canonical_record":{"source":{"id":"2408.00761","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-08-01T17:59:12Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"70573cc4b6dce644ec0c38f18f643a2c9ff0ed9e6e1c1d1e692c9a4e1c629d1b","abstract_canon_sha256":"0fe46027c91b0e8578b6e6461d93a8b2f0631a94672f975a9e3fff27e27ec259"},"schema_version":"1.0"},"canonical_sha256":"04a56730db8d866e571cc5573e1ff5155ef2321465a443eb5fa9dbc7c42d1449","source":{"kind":"arxiv","id":"2408.00761","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2408.00761","created_at":"2026-07-05T10:12:00Z"},{"alias_kind":"arxiv_version","alias_value":"2408.00761v4","created_at":"2026-07-05T10:12:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.00761","created_at":"2026-07-05T10:12:00Z"},{"alias_kind":"pith_short_12","alias_value":"ASSWOMG3RWDG","created_at":"2026-07-05T10:12:00Z"},{"alias_kind":"pith_short_16","alias_value":"ASSWOMG3RWDG4VY4","created_at":"2026-07-05T10:12:00Z"},{"alias_kind":"pith_short_8","alias_value":"ASSWOMG3","created_at":"2026-07-05T10:12:00Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:ASSWOMG3RWDG4VY4YVLT4H7VCV","target":"record","payload":{"canonical_record":{"source":{"id":"2408.00761","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-08-01T17:59:12Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"70573cc4b6dce644ec0c38f18f643a2c9ff0ed9e6e1c1d1e692c9a4e1c629d1b","abstract_canon_sha256":"0fe46027c91b0e8578b6e6461d93a8b2f0631a94672f975a9e3fff27e27ec259"},"schema_version":"1.0"},"canonical_sha256":"04a56730db8d866e571cc5573e1ff5155ef2321465a443eb5fa9dbc7c42d1449","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:12:00.270430Z","signature_b64":"xE/NXLTtsVDVqH3hUDBW9qmGIonPF8lVU1v110pqaA8DgeFnIBwTJtIi1vChneY+1CvqbZXCmMS7YPX1h42ECg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"04a56730db8d866e571cc5573e1ff5155ef2321465a443eb5fa9dbc7c42d1449","last_reissued_at":"2026-07-05T10:12:00.269902Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:12:00.269902Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2408.00761","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":"2026-07-05T10:12:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"9Pj81V5YJK7BqYUK4oiAPBy1GrNtqO7XpHJSfjTx7FQn2pXMBSMXdTWJ0c1k7P991xhfAWNZzVIw2mTIzSU0BQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T20:44:57.272357Z"},"content_sha256":"526bb050505ec93003439bfb47dd42efa0ba03e5572820095dee663121148310","schema_version":"1.0","event_id":"sha256:526bb050505ec93003439bfb47dd42efa0ba03e5572820095dee663121148310"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:ASSWOMG3RWDG4VY4YVLT4H7VCV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Tamper-Resistant Safeguards for Open-Weight LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Alice Gatti, Andy Zhou, Andy Zou, Bhrugu Bharathi, Bo Li, Dan Hendrycks, Dawn Song, Justin Wang, Long Phan, Mantas Mazeika, Maxwell Lin, Rishub Tamirisa, Ron Arel, Rowan Wang, Tarun Suresh","submitted_at":"2024-08-01T17:59:12Z","abstract_excerpt":"Rapid advances in the capabilities of large language models (LLMs) have raised widespread concerns regarding their potential for malicious use. Open-weight LLMs present unique challenges, as existing safeguards lack robustness to tampering attacks that modify model weights. For example, recent works have demonstrated that refusal and unlearning safeguards can be trivially removed with a few steps of fine-tuning. These vulnerabilities necessitate new approaches for enabling the safe release of open-weight LLMs. We develop a method, called TAR, for building tamper-resistant safeguards into open-"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.00761","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/2408.00761/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":"2026-07-05T10:12:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"vlsxfs7+dHoJQDKku6ejC1B/HiMglGnAKI+4koxwV3NuKyssI2aEOylWCebdm2sxEefpIJXRZgvEVyaBNe7eBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T20:44:57.273209Z"},"content_sha256":"68fb2f03666dd51b7759eb0b659ff65ae8bd5473db804505f613c3fb9e79a606","schema_version":"1.0","event_id":"sha256:68fb2f03666dd51b7759eb0b659ff65ae8bd5473db804505f613c3fb9e79a606"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/ASSWOMG3RWDG4VY4YVLT4H7VCV/bundle.json","state_url":"https://pith.science/pith/ASSWOMG3RWDG4VY4YVLT4H7VCV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/ASSWOMG3RWDG4VY4YVLT4H7VCV/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":"2026-08-05T20:44:57Z","links":{"resolver":"https://pith.science/pith/ASSWOMG3RWDG4VY4YVLT4H7VCV","bundle":"https://pith.science/pith/ASSWOMG3RWDG4VY4YVLT4H7VCV/bundle.json","state":"https://pith.science/pith/ASSWOMG3RWDG4VY4YVLT4H7VCV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/ASSWOMG3RWDG4VY4YVLT4H7VCV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:ASSWOMG3RWDG4VY4YVLT4H7VCV","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":"0fe46027c91b0e8578b6e6461d93a8b2f0631a94672f975a9e3fff27e27ec259","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-08-01T17:59:12Z","title_canon_sha256":"70573cc4b6dce644ec0c38f18f643a2c9ff0ed9e6e1c1d1e692c9a4e1c629d1b"},"schema_version":"1.0","source":{"id":"2408.00761","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2408.00761","created_at":"2026-07-05T10:12:00Z"},{"alias_kind":"arxiv_version","alias_value":"2408.00761v4","created_at":"2026-07-05T10:12:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.00761","created_at":"2026-07-05T10:12:00Z"},{"alias_kind":"pith_short_12","alias_value":"ASSWOMG3RWDG","created_at":"2026-07-05T10:12:00Z"},{"alias_kind":"pith_short_16","alias_value":"ASSWOMG3RWDG4VY4","created_at":"2026-07-05T10:12:00Z"},{"alias_kind":"pith_short_8","alias_value":"ASSWOMG3","created_at":"2026-07-05T10:12:00Z"}],"graph_snapshots":[{"event_id":"sha256:68fb2f03666dd51b7759eb0b659ff65ae8bd5473db804505f613c3fb9e79a606","target":"graph","created_at":"2026-07-05T10:12:00Z","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/2408.00761/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Rapid advances in the capabilities of large language models (LLMs) have raised widespread concerns regarding their potential for malicious use. Open-weight LLMs present unique challenges, as existing safeguards lack robustness to tampering attacks that modify model weights. For example, recent works have demonstrated that refusal and unlearning safeguards can be trivially removed with a few steps of fine-tuning. These vulnerabilities necessitate new approaches for enabling the safe release of open-weight LLMs. We develop a method, called TAR, for building tamper-resistant safeguards into open-","authors_text":"Alice Gatti, Andy Zhou, Andy Zou, Bhrugu Bharathi, Bo Li, Dan Hendrycks, Dawn Song, Justin Wang, Long Phan, Mantas Mazeika, Maxwell Lin, Rishub Tamirisa, Ron Arel, Rowan Wang, Tarun Suresh","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-08-01T17:59:12Z","title":"Tamper-Resistant Safeguards for Open-Weight LLMs"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.00761","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":"sha256:526bb050505ec93003439bfb47dd42efa0ba03e5572820095dee663121148310","target":"record","created_at":"2026-07-05T10:12:00Z","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":"0fe46027c91b0e8578b6e6461d93a8b2f0631a94672f975a9e3fff27e27ec259","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-08-01T17:59:12Z","title_canon_sha256":"70573cc4b6dce644ec0c38f18f643a2c9ff0ed9e6e1c1d1e692c9a4e1c629d1b"},"schema_version":"1.0","source":{"id":"2408.00761","kind":"arxiv","version":4}},"canonical_sha256":"04a56730db8d866e571cc5573e1ff5155ef2321465a443eb5fa9dbc7c42d1449","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"04a56730db8d866e571cc5573e1ff5155ef2321465a443eb5fa9dbc7c42d1449","first_computed_at":"2026-07-05T10:12:00.269902Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:12:00.269902Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"xE/NXLTtsVDVqH3hUDBW9qmGIonPF8lVU1v110pqaA8DgeFnIBwTJtIi1vChneY+1CvqbZXCmMS7YPX1h42ECg==","signature_status":"signed_v1","signed_at":"2026-07-05T10:12:00.270430Z","signed_message":"canonical_sha256_bytes"},"source_id":"2408.00761","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:526bb050505ec93003439bfb47dd42efa0ba03e5572820095dee663121148310","sha256:68fb2f03666dd51b7759eb0b659ff65ae8bd5473db804505f613c3fb9e79a606"],"state_sha256":"5ea366e161ace4ac10b7005404a40d7467ce06f7d918840eb4f26ffa6510994b"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"fZ3qCiMSljFUJD1u1gQLA9AvaDgJNZB7sFfEezadyow77+mpJbVhfLJGp+9oauh2BlUJHfUMGimsW6+RGwQdBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-05T20:44:57.280917Z","bundle_sha256":"1ca82daca7ce5dff6cf9bc01c48853675cbf47ba6a70338cf9e17f77dbf7a175"}}