{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:ZHV6GX2556ABJWK2KIA5VXRMTS","short_pith_number":"pith:ZHV6GX25","canonical_record":{"source":{"id":"2405.18166","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-05-28T13:26:12Z","cross_cats_sorted":[],"title_canon_sha256":"ff1efc823a5f3c1af5b5964f6113277d674970d2aa5aa17e603dad74ce80d2e5","abstract_canon_sha256":"ef54ef1ee622f8f845db8a938babbcfdd5f88fb39397ef1ca59ec3c4943f1735"},"schema_version":"1.0"},"canonical_sha256":"c9ebe35f5def8014d95a5201dade2c9cb65c075ea822c774dbc72c07876a970b","source":{"kind":"arxiv","id":"2405.18166","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.18166","created_at":"2026-07-05T08:31:51Z"},{"alias_kind":"arxiv_version","alias_value":"2405.18166v2","created_at":"2026-07-05T08:31:51Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.18166","created_at":"2026-07-05T08:31:51Z"},{"alias_kind":"pith_short_12","alias_value":"ZHV6GX2556AB","created_at":"2026-07-05T08:31:51Z"},{"alias_kind":"pith_short_16","alias_value":"ZHV6GX2556ABJWK2","created_at":"2026-07-05T08:31:51Z"},{"alias_kind":"pith_short_8","alias_value":"ZHV6GX25","created_at":"2026-07-05T08:31:51Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:ZHV6GX2556ABJWK2KIA5VXRMTS","target":"record","payload":{"canonical_record":{"source":{"id":"2405.18166","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-05-28T13:26:12Z","cross_cats_sorted":[],"title_canon_sha256":"ff1efc823a5f3c1af5b5964f6113277d674970d2aa5aa17e603dad74ce80d2e5","abstract_canon_sha256":"ef54ef1ee622f8f845db8a938babbcfdd5f88fb39397ef1ca59ec3c4943f1735"},"schema_version":"1.0"},"canonical_sha256":"c9ebe35f5def8014d95a5201dade2c9cb65c075ea822c774dbc72c07876a970b","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:31:51.618919Z","signature_b64":"PgArI8emntHMuoVMoKtPlLeniwq7qQHh33slGg3ZH8qPOQIvabC3vvXwfavoRWi2rfd7QCVH88/821+D6dNbBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c9ebe35f5def8014d95a5201dade2c9cb65c075ea822c774dbc72c07876a970b","last_reissued_at":"2026-07-05T08:31:51.618421Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:31:51.618421Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.18166","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":"2026-07-05T08:31:51Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"tuWjfhYw79qVdiFwHpqNMqqs09hoZuDTqRnM6S1UCSn4HTcgVLa7qurxzTVrSfIqhdNlQLjb4E3ai7xY3lFOBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T08:21:43.472384Z"},"content_sha256":"283e1f5145135bc53261f53e5ea8ec268eb164e741b45778ca5471abe3f74e1e","schema_version":"1.0","event_id":"sha256:283e1f5145135bc53261f53e5ea8ec268eb164e741b45778ca5471abe3f74e1e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:ZHV6GX2556ABJWK2KIA5VXRMTS","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Defending Large Language Models Against Jailbreak Attacks via Layer-specific Editing","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jun Sun, Wei Zhao, Ye Zhang, Yige Li, Zhe Li","submitted_at":"2024-05-28T13:26:12Z","abstract_excerpt":"Large language models (LLMs) are increasingly being adopted in a wide range of real-world applications. Despite their impressive performance, recent studies have shown that LLMs are vulnerable to deliberately crafted adversarial prompts even when aligned via Reinforcement Learning from Human Feedback or supervised fine-tuning. While existing defense methods focus on either detecting harmful prompts or reducing the likelihood of harmful responses through various means, defending LLMs against jailbreak attacks based on the inner mechanisms of LLMs remains largely unexplored. In this work, we inv"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.18166","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/2405.18166/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":"2026-07-05T08:31:51Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"eUfyATeTTL2qYUAflurf4+Nm60iw0+Nf3/tBSvP1WVDP+ZUcH5FW20nZho2/eIdL5nWG62S+aSXPSj7vWT4/Aw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T08:21:43.472968Z"},"content_sha256":"630e1f3e4cbb6ceec9beb6d132ad82d48bbafe5949e1af48052739f47c920d1d","schema_version":"1.0","event_id":"sha256:630e1f3e4cbb6ceec9beb6d132ad82d48bbafe5949e1af48052739f47c920d1d"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/ZHV6GX2556ABJWK2KIA5VXRMTS/bundle.json","state_url":"https://pith.science/pith/ZHV6GX2556ABJWK2KIA5VXRMTS/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/ZHV6GX2556ABJWK2KIA5VXRMTS/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":"2026-08-09T08:21:43Z","links":{"resolver":"https://pith.science/pith/ZHV6GX2556ABJWK2KIA5VXRMTS","bundle":"https://pith.science/pith/ZHV6GX2556ABJWK2KIA5VXRMTS/bundle.json","state":"https://pith.science/pith/ZHV6GX2556ABJWK2KIA5VXRMTS/state.json","well_known_bundle":"https://pith.science/.well-known/pith/ZHV6GX2556ABJWK2KIA5VXRMTS/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:ZHV6GX2556ABJWK2KIA5VXRMTS","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":"ef54ef1ee622f8f845db8a938babbcfdd5f88fb39397ef1ca59ec3c4943f1735","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-05-28T13:26:12Z","title_canon_sha256":"ff1efc823a5f3c1af5b5964f6113277d674970d2aa5aa17e603dad74ce80d2e5"},"schema_version":"1.0","source":{"id":"2405.18166","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.18166","created_at":"2026-07-05T08:31:51Z"},{"alias_kind":"arxiv_version","alias_value":"2405.18166v2","created_at":"2026-07-05T08:31:51Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.18166","created_at":"2026-07-05T08:31:51Z"},{"alias_kind":"pith_short_12","alias_value":"ZHV6GX2556AB","created_at":"2026-07-05T08:31:51Z"},{"alias_kind":"pith_short_16","alias_value":"ZHV6GX2556ABJWK2","created_at":"2026-07-05T08:31:51Z"},{"alias_kind":"pith_short_8","alias_value":"ZHV6GX25","created_at":"2026-07-05T08:31:51Z"}],"graph_snapshots":[{"event_id":"sha256:630e1f3e4cbb6ceec9beb6d132ad82d48bbafe5949e1af48052739f47c920d1d","target":"graph","created_at":"2026-07-05T08:31:51Z","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/2405.18166/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) are increasingly being adopted in a wide range of real-world applications. Despite their impressive performance, recent studies have shown that LLMs are vulnerable to deliberately crafted adversarial prompts even when aligned via Reinforcement Learning from Human Feedback or supervised fine-tuning. While existing defense methods focus on either detecting harmful prompts or reducing the likelihood of harmful responses through various means, defending LLMs against jailbreak attacks based on the inner mechanisms of LLMs remains largely unexplored. In this work, we inv","authors_text":"Jun Sun, Wei Zhao, Ye Zhang, Yige Li, Zhe Li","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-05-28T13:26:12Z","title":"Defending Large Language Models Against Jailbreak Attacks via Layer-specific Editing"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.18166","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":"sha256:283e1f5145135bc53261f53e5ea8ec268eb164e741b45778ca5471abe3f74e1e","target":"record","created_at":"2026-07-05T08:31:51Z","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":"ef54ef1ee622f8f845db8a938babbcfdd5f88fb39397ef1ca59ec3c4943f1735","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-05-28T13:26:12Z","title_canon_sha256":"ff1efc823a5f3c1af5b5964f6113277d674970d2aa5aa17e603dad74ce80d2e5"},"schema_version":"1.0","source":{"id":"2405.18166","kind":"arxiv","version":2}},"canonical_sha256":"c9ebe35f5def8014d95a5201dade2c9cb65c075ea822c774dbc72c07876a970b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c9ebe35f5def8014d95a5201dade2c9cb65c075ea822c774dbc72c07876a970b","first_computed_at":"2026-07-05T08:31:51.618421Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:31:51.618421Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"PgArI8emntHMuoVMoKtPlLeniwq7qQHh33slGg3ZH8qPOQIvabC3vvXwfavoRWi2rfd7QCVH88/821+D6dNbBQ==","signature_status":"signed_v1","signed_at":"2026-07-05T08:31:51.618919Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.18166","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:283e1f5145135bc53261f53e5ea8ec268eb164e741b45778ca5471abe3f74e1e","sha256:630e1f3e4cbb6ceec9beb6d132ad82d48bbafe5949e1af48052739f47c920d1d"],"state_sha256":"40502b7f678e69c3cbba64d11288e35c2928124b26296f7261e13e712ddbe6b8"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"mD8b3gO/VWF+QW02aw5YfkwvfURpoUwlFlHqz6z3YgpMYdZ0yDhn7HzsH2lQ2++odAwZ5SSOA+oYBc2vjWRSAQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T08:21:43.478505Z","bundle_sha256":"38fa9dd9fb227f23465813c1e770441496aa827e251cdfdac7259713062975d0"}}