{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:JXI5SPIP2BBF7DUR6NORLUEBZP","short_pith_number":"pith:JXI5SPIP","canonical_record":{"source":{"id":"2406.17092","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-06-24T19:29:47Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"bf0ecd1f8e6aaa68c5e4274362dc04141627f2a0310b8bd9bbc1656aa0d80846","abstract_canon_sha256":"3bc707f55183160a888e3d493f195eca831e8136a9de1ceca988bec437548017"},"schema_version":"1.0"},"canonical_sha256":"4dd1d93d0fd0425f8e91f35d15d081cbeba6acb71415cc3417c09821f86bdab1","source":{"kind":"arxiv","id":"2406.17092","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.17092","created_at":"2026-07-05T08:36:09Z"},{"alias_kind":"arxiv_version","alias_value":"2406.17092v1","created_at":"2026-07-05T08:36:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.17092","created_at":"2026-07-05T08:36:09Z"},{"alias_kind":"pith_short_12","alias_value":"JXI5SPIP2BBF","created_at":"2026-07-05T08:36:09Z"},{"alias_kind":"pith_short_16","alias_value":"JXI5SPIP2BBF7DUR","created_at":"2026-07-05T08:36:09Z"},{"alias_kind":"pith_short_8","alias_value":"JXI5SPIP","created_at":"2026-07-05T08:36:09Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:JXI5SPIP2BBF7DUR6NORLUEBZP","target":"record","payload":{"canonical_record":{"source":{"id":"2406.17092","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-06-24T19:29:47Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"bf0ecd1f8e6aaa68c5e4274362dc04141627f2a0310b8bd9bbc1656aa0d80846","abstract_canon_sha256":"3bc707f55183160a888e3d493f195eca831e8136a9de1ceca988bec437548017"},"schema_version":"1.0"},"canonical_sha256":"4dd1d93d0fd0425f8e91f35d15d081cbeba6acb71415cc3417c09821f86bdab1","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:36:09.974148Z","signature_b64":"zuRi1CF6Y4mOjFFjq1bepQxME4eAcNOJsog44iQB1/zb4spKuQ3hvDVQudZlnKZpso+3Bb7qrzRyhoMpL4tnCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4dd1d93d0fd0425f8e91f35d15d081cbeba6acb71415cc3417c09821f86bdab1","last_reissued_at":"2026-07-05T08:36:09.973739Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:36:09.973739Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2406.17092","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:36:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"I23tnkxhRDsiL85qsfXgoTtDsTWrTPyR9+oyn8X0lkLoCBuprUb0mbFbDVBuLXX1oSGXznqvIEv+c9QpPapcBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T23:26:10.793369Z"},"content_sha256":"77c761c08bfaa44bd552344825bb2ed1739cf9a92e874291f30c755065d84afd","schema_version":"1.0","event_id":"sha256:77c761c08bfaa44bd552344825bb2ed1739cf9a92e874291f30c755065d84afd"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:JXI5SPIP2BBF7DUR6NORLUEBZP","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"BEEAR: Embedding-based Adversarial Removal of Safety Backdoors in Instruction-tuned Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Bo Li, Dawn Song, Ruoxi Jia, Tran Ngoc Huynh, Weiyu Sun, Yi Zeng","submitted_at":"2024-06-24T19:29:47Z","abstract_excerpt":"Safety backdoor attacks in large language models (LLMs) enable the stealthy triggering of unsafe behaviors while evading detection during normal interactions. The high dimensionality of potential triggers in the token space and the diverse range of malicious behaviors make this a critical challenge. We present BEEAR, a mitigation approach leveraging the insight that backdoor triggers induce relatively uniform drifts in the model's embedding space. Our bi-level optimization method identifies universal embedding perturbations that elicit unwanted behaviors and adjusts the model parameters to rei"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.17092","kind":"arxiv","version":1},"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/2406.17092/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":"2026-07-05T08:36:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"eQbRScOGXVooLGvAgc8EUGn9SQPZouFlw73cLacM4y0xE5ekC5mqypq6/1AjUoi/lYMtWFu0835SBWwFevVBBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T23:26:10.793955Z"},"content_sha256":"82d79f60f96841421b39de92ecefd8b64e0c68f9ff7a57fcfe5957525bcc5e22","schema_version":"1.0","event_id":"sha256:82d79f60f96841421b39de92ecefd8b64e0c68f9ff7a57fcfe5957525bcc5e22"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/JXI5SPIP2BBF7DUR6NORLUEBZP/bundle.json","state_url":"https://pith.science/pith/JXI5SPIP2BBF7DUR6NORLUEBZP/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/JXI5SPIP2BBF7DUR6NORLUEBZP/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":"2026-08-08T23:26:10Z","links":{"resolver":"https://pith.science/pith/JXI5SPIP2BBF7DUR6NORLUEBZP","bundle":"https://pith.science/pith/JXI5SPIP2BBF7DUR6NORLUEBZP/bundle.json","state":"https://pith.science/pith/JXI5SPIP2BBF7DUR6NORLUEBZP/state.json","well_known_bundle":"https://pith.science/.well-known/pith/JXI5SPIP2BBF7DUR6NORLUEBZP/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:JXI5SPIP2BBF7DUR6NORLUEBZP","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":"3bc707f55183160a888e3d493f195eca831e8136a9de1ceca988bec437548017","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-06-24T19:29:47Z","title_canon_sha256":"bf0ecd1f8e6aaa68c5e4274362dc04141627f2a0310b8bd9bbc1656aa0d80846"},"schema_version":"1.0","source":{"id":"2406.17092","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.17092","created_at":"2026-07-05T08:36:09Z"},{"alias_kind":"arxiv_version","alias_value":"2406.17092v1","created_at":"2026-07-05T08:36:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.17092","created_at":"2026-07-05T08:36:09Z"},{"alias_kind":"pith_short_12","alias_value":"JXI5SPIP2BBF","created_at":"2026-07-05T08:36:09Z"},{"alias_kind":"pith_short_16","alias_value":"JXI5SPIP2BBF7DUR","created_at":"2026-07-05T08:36:09Z"},{"alias_kind":"pith_short_8","alias_value":"JXI5SPIP","created_at":"2026-07-05T08:36:09Z"}],"graph_snapshots":[{"event_id":"sha256:82d79f60f96841421b39de92ecefd8b64e0c68f9ff7a57fcfe5957525bcc5e22","target":"graph","created_at":"2026-07-05T08:36:09Z","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/2406.17092/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Safety backdoor attacks in large language models (LLMs) enable the stealthy triggering of unsafe behaviors while evading detection during normal interactions. The high dimensionality of potential triggers in the token space and the diverse range of malicious behaviors make this a critical challenge. We present BEEAR, a mitigation approach leveraging the insight that backdoor triggers induce relatively uniform drifts in the model's embedding space. Our bi-level optimization method identifies universal embedding perturbations that elicit unwanted behaviors and adjusts the model parameters to rei","authors_text":"Bo Li, Dawn Song, Ruoxi Jia, Tran Ngoc Huynh, Weiyu Sun, Yi Zeng","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-06-24T19:29:47Z","title":"BEEAR: Embedding-based Adversarial Removal of Safety Backdoors in Instruction-tuned Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.17092","kind":"arxiv","version":1},"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":"sha256:77c761c08bfaa44bd552344825bb2ed1739cf9a92e874291f30c755065d84afd","target":"record","created_at":"2026-07-05T08:36:09Z","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":"3bc707f55183160a888e3d493f195eca831e8136a9de1ceca988bec437548017","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-06-24T19:29:47Z","title_canon_sha256":"bf0ecd1f8e6aaa68c5e4274362dc04141627f2a0310b8bd9bbc1656aa0d80846"},"schema_version":"1.0","source":{"id":"2406.17092","kind":"arxiv","version":1}},"canonical_sha256":"4dd1d93d0fd0425f8e91f35d15d081cbeba6acb71415cc3417c09821f86bdab1","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4dd1d93d0fd0425f8e91f35d15d081cbeba6acb71415cc3417c09821f86bdab1","first_computed_at":"2026-07-05T08:36:09.973739Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:36:09.973739Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"zuRi1CF6Y4mOjFFjq1bepQxME4eAcNOJsog44iQB1/zb4spKuQ3hvDVQudZlnKZpso+3Bb7qrzRyhoMpL4tnCg==","signature_status":"signed_v1","signed_at":"2026-07-05T08:36:09.974148Z","signed_message":"canonical_sha256_bytes"},"source_id":"2406.17092","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:77c761c08bfaa44bd552344825bb2ed1739cf9a92e874291f30c755065d84afd","sha256:82d79f60f96841421b39de92ecefd8b64e0c68f9ff7a57fcfe5957525bcc5e22"],"state_sha256":"c263e71129ba11c3415f0030d7cb5d07d4dff51ced369b29172bf8dc8db0c357"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"6pJVR2PIoIVm+Fq6OoE53nrRAgfxD42/W2poJ4beH5kSi2Xf1GIgLUeis4jA1CJZjCEi5eTMMt1/w95E+37KCw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T23:26:10.798758Z","bundle_sha256":"b703d6ff9d54ae92ef208b71c6fcd2fa9004563d5e2d735625b4037878b31233"}}