{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:4MTKL7CL53P4OA7STXULTIO6KV","short_pith_number":"pith:4MTKL7CL","canonical_record":{"source":{"id":"2405.13820","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-22T16:51:07Z","cross_cats_sorted":[],"title_canon_sha256":"b5ea66ea0da5e85128fe9172ea69ec7635fddf8980addcfd1e7767deef128b00","abstract_canon_sha256":"d208a78f745035370726c25082dc6401957e07d7c203b998536243dc5db75642"},"schema_version":"1.0"},"canonical_sha256":"e326a5fc4beedfc703f29de8b9a1de555a2c83ec45a2fa8c8d603a778db12dae","source":{"kind":"arxiv","id":"2405.13820","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.13820","created_at":"2026-07-05T09:50:02Z"},{"alias_kind":"arxiv_version","alias_value":"2405.13820v2","created_at":"2026-07-05T09:50:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.13820","created_at":"2026-07-05T09:50:02Z"},{"alias_kind":"pith_short_12","alias_value":"4MTKL7CL53P4","created_at":"2026-07-05T09:50:02Z"},{"alias_kind":"pith_short_16","alias_value":"4MTKL7CL53P4OA7S","created_at":"2026-07-05T09:50:02Z"},{"alias_kind":"pith_short_8","alias_value":"4MTKL7CL","created_at":"2026-07-05T09:50:02Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:4MTKL7CL53P4OA7STXULTIO6KV","target":"record","payload":{"canonical_record":{"source":{"id":"2405.13820","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-22T16:51:07Z","cross_cats_sorted":[],"title_canon_sha256":"b5ea66ea0da5e85128fe9172ea69ec7635fddf8980addcfd1e7767deef128b00","abstract_canon_sha256":"d208a78f745035370726c25082dc6401957e07d7c203b998536243dc5db75642"},"schema_version":"1.0"},"canonical_sha256":"e326a5fc4beedfc703f29de8b9a1de555a2c83ec45a2fa8c8d603a778db12dae","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:50:02.588092Z","signature_b64":"625dBwaeLXYTmIk1fLHOAGh5HjwLT6uN3i9KbeyuUMOeNsIzTX2iDiUSdI486zQ4U+81CrCeumB9ZnJ2ZRhWBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e326a5fc4beedfc703f29de8b9a1de555a2c83ec45a2fa8c8d603a778db12dae","last_reissued_at":"2026-07-05T09:50:02.587607Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:50:02.587607Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.13820","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:50:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"s8J/gfPo3Apk2zdq/PEhqVAbAiaAqUNQxaIuSyJkWCbSXoUi2fIhrYJOQGOoADCWT7p71QmI3wm+6B7TVDW+CQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-09T02:36:52.998128Z"},"content_sha256":"f571a21715768a81a43598a4f0748e14eb1bcce1d1c4e5a6d7b02ace0156f062","schema_version":"1.0","event_id":"sha256:f571a21715768a81a43598a4f0748e14eb1bcce1d1c4e5a6d7b02ace0156f062"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:4MTKL7CL53P4OA7STXULTIO6KV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Towards Comprehensive Post Safety Alignment of Large Language Models via Safety Patching","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bing Qin, Jiahe Guo, Tat-Seng Chua, Ting Liu, Weixiang Zhao, Xingyu Sui, Yang Deng, Yanyan Zhao, Yulin Hu, Zhuojun Li","submitted_at":"2024-05-22T16:51:07Z","abstract_excerpt":"Safety alignment of large language models (LLMs) has been gaining increasing attention. However, current safety-aligned LLMs suffer from the fragile and imbalanced safety mechanisms, which can still be induced to generate unsafe responses, exhibit over-safety by rejecting safe user inputs, and fail to preserve general utility after safety alignment. To this end, we propose a novel post safety alignment (PSA) method to address these inherent and emerging safety challenges, including safety enhancement, over-safety mitigation, and utility preservation. In specific, we introduce \\textsc{SafePatch"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.13820","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/2405.13820/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":"2026-07-05T09:50:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"175VbeOzXGNO4JYaMF9NbqNDla2ZuERYB2uWcf5WzTURUEdwnrgVRrzggEgR5N8BHnNLm/5x3Y3WsKl3PlHoAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-09T02:36:52.998580Z"},"content_sha256":"faa1cdc73c7323b8322950d820cd60ac5a162d896c5ea3ee9407b7055becdaf0","schema_version":"1.0","event_id":"sha256:faa1cdc73c7323b8322950d820cd60ac5a162d896c5ea3ee9407b7055becdaf0"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/4MTKL7CL53P4OA7STXULTIO6KV/bundle.json","state_url":"https://pith.science/pith/4MTKL7CL53P4OA7STXULTIO6KV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/4MTKL7CL53P4OA7STXULTIO6KV/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":"2026-07-09T02:36:52Z","links":{"resolver":"https://pith.science/pith/4MTKL7CL53P4OA7STXULTIO6KV","bundle":"https://pith.science/pith/4MTKL7CL53P4OA7STXULTIO6KV/bundle.json","state":"https://pith.science/pith/4MTKL7CL53P4OA7STXULTIO6KV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/4MTKL7CL53P4OA7STXULTIO6KV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:4MTKL7CL53P4OA7STXULTIO6KV","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":"d208a78f745035370726c25082dc6401957e07d7c203b998536243dc5db75642","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-22T16:51:07Z","title_canon_sha256":"b5ea66ea0da5e85128fe9172ea69ec7635fddf8980addcfd1e7767deef128b00"},"schema_version":"1.0","source":{"id":"2405.13820","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.13820","created_at":"2026-07-05T09:50:02Z"},{"alias_kind":"arxiv_version","alias_value":"2405.13820v2","created_at":"2026-07-05T09:50:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.13820","created_at":"2026-07-05T09:50:02Z"},{"alias_kind":"pith_short_12","alias_value":"4MTKL7CL53P4","created_at":"2026-07-05T09:50:02Z"},{"alias_kind":"pith_short_16","alias_value":"4MTKL7CL53P4OA7S","created_at":"2026-07-05T09:50:02Z"},{"alias_kind":"pith_short_8","alias_value":"4MTKL7CL","created_at":"2026-07-05T09:50:02Z"}],"graph_snapshots":[{"event_id":"sha256:faa1cdc73c7323b8322950d820cd60ac5a162d896c5ea3ee9407b7055becdaf0","target":"graph","created_at":"2026-07-05T09:50:02Z","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/2405.13820/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Safety alignment of large language models (LLMs) has been gaining increasing attention. However, current safety-aligned LLMs suffer from the fragile and imbalanced safety mechanisms, which can still be induced to generate unsafe responses, exhibit over-safety by rejecting safe user inputs, and fail to preserve general utility after safety alignment. To this end, we propose a novel post safety alignment (PSA) method to address these inherent and emerging safety challenges, including safety enhancement, over-safety mitigation, and utility preservation. In specific, we introduce \\textsc{SafePatch","authors_text":"Bing Qin, Jiahe Guo, Tat-Seng Chua, Ting Liu, Weixiang Zhao, Xingyu Sui, Yang Deng, Yanyan Zhao, Yulin Hu, Zhuojun Li","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-22T16:51:07Z","title":"Towards Comprehensive Post Safety Alignment of Large Language Models via Safety Patching"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.13820","kind":"arxiv","version":2},"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:f571a21715768a81a43598a4f0748e14eb1bcce1d1c4e5a6d7b02ace0156f062","target":"record","created_at":"2026-07-05T09:50:02Z","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":"d208a78f745035370726c25082dc6401957e07d7c203b998536243dc5db75642","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-22T16:51:07Z","title_canon_sha256":"b5ea66ea0da5e85128fe9172ea69ec7635fddf8980addcfd1e7767deef128b00"},"schema_version":"1.0","source":{"id":"2405.13820","kind":"arxiv","version":2}},"canonical_sha256":"e326a5fc4beedfc703f29de8b9a1de555a2c83ec45a2fa8c8d603a778db12dae","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e326a5fc4beedfc703f29de8b9a1de555a2c83ec45a2fa8c8d603a778db12dae","first_computed_at":"2026-07-05T09:50:02.587607Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:50:02.587607Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"625dBwaeLXYTmIk1fLHOAGh5HjwLT6uN3i9KbeyuUMOeNsIzTX2iDiUSdI486zQ4U+81CrCeumB9ZnJ2ZRhWBg==","signature_status":"signed_v1","signed_at":"2026-07-05T09:50:02.588092Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.13820","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f571a21715768a81a43598a4f0748e14eb1bcce1d1c4e5a6d7b02ace0156f062","sha256:faa1cdc73c7323b8322950d820cd60ac5a162d896c5ea3ee9407b7055becdaf0"],"state_sha256":"552162f6d57d9e11d4579ad6e7c075c7517eb44b4eac5ce125d3d5a708e2537b"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3InICM1bls1HVPF7RwAhvlBo7+2Y+ilWXVyq3v6y7YUzKr5y6ClLPFSlYPOwiTqCwOnXDERzAUkfw+otbX/BBg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-09T02:36:53.000874Z","bundle_sha256":"1414ee0ee28c7b7f64adf4a121e552253cdca589119d06eefa0944370ecb8a75"}}