{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:26ZTRR2F22FANJVBUOTQVLTJMI","short_pith_number":"pith:26ZTRR2F","canonical_record":{"source":{"id":"2404.05880","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-08T21:26:22Z","cross_cats_sorted":[],"title_canon_sha256":"ea0b933138ca0480367ac48901ae492f3871422db40a5c9ac676aeee2cc33106","abstract_canon_sha256":"ea311be10a3a1c437b7236e3a493e6b61f64b51aa558ce9d62b128e35483f1c6"},"schema_version":"1.0"},"canonical_sha256":"d7b338c745d68a06a6a1a3a70aae69622808f9c2c73f6042c6469a6ed110767e","source":{"kind":"arxiv","id":"2404.05880","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.05880","created_at":"2026-07-05T08:39:33Z"},{"alias_kind":"arxiv_version","alias_value":"2404.05880v2","created_at":"2026-07-05T08:39:33Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.05880","created_at":"2026-07-05T08:39:33Z"},{"alias_kind":"pith_short_12","alias_value":"26ZTRR2F22FA","created_at":"2026-07-05T08:39:33Z"},{"alias_kind":"pith_short_16","alias_value":"26ZTRR2F22FANJVB","created_at":"2026-07-05T08:39:33Z"},{"alias_kind":"pith_short_8","alias_value":"26ZTRR2F","created_at":"2026-07-05T08:39:33Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:26ZTRR2F22FANJVBUOTQVLTJMI","target":"record","payload":{"canonical_record":{"source":{"id":"2404.05880","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-08T21:26:22Z","cross_cats_sorted":[],"title_canon_sha256":"ea0b933138ca0480367ac48901ae492f3871422db40a5c9ac676aeee2cc33106","abstract_canon_sha256":"ea311be10a3a1c437b7236e3a493e6b61f64b51aa558ce9d62b128e35483f1c6"},"schema_version":"1.0"},"canonical_sha256":"d7b338c745d68a06a6a1a3a70aae69622808f9c2c73f6042c6469a6ed110767e","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:39:33.340433Z","signature_b64":"iaasHWwIJWEesviTKgm0eBB/6j4+AXCfRHKQ6P8EEsv12HvHkhAg/bddMNIFQDbEIPPfFktnrf6Y5epXVygLBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d7b338c745d68a06a6a1a3a70aae69622808f9c2c73f6042c6469a6ed110767e","last_reissued_at":"2026-07-05T08:39:33.339945Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:39:33.339945Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2404.05880","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":"2026-07-05T08:39:33Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"HuWayJh23UDj6PhAagWjUTzDJNV/EWYA957aD8kyRmL7QV88c7juN4l2FEC5DauE/gPkuvC04w1u33lFCUxzDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T18:34:33.288345Z"},"content_sha256":"a78ca18a964a35f55204091870495daea21cc9a47a3dab1f114ed696286e3aa4","schema_version":"1.0","event_id":"sha256:a78ca18a964a35f55204091870495daea21cc9a47a3dab1f114ed696286e3aa4"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:26ZTRR2F22FANJVBUOTQVLTJMI","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Eraser: Jailbreaking Defense in Large Language Models via Unlearning Harmful Knowledge","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Cen Chen, Huiping Zhuang, Jianwei Wang, Weikai Lu, Zelin Chen, Zhengdong Lu, Ziqian Zeng","submitted_at":"2024-04-08T21:26:22Z","abstract_excerpt":"Jailbreaking attacks can enable Large Language Models (LLMs) to bypass the safeguard and generate harmful content. Existing jailbreaking defense methods have failed to address the fundamental issue that harmful knowledge resides within the model, leading to potential jailbreak risks for LLMs. In this paper, we propose a novel defense method called Eraser, which mainly includes three goals: unlearning harmful knowledge, retaining general knowledge, and maintaining safety alignment. The intuition is that if an LLM forgets the specific knowledge required to answer a harmful question, it will no l"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.05880","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/2404.05880/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":"2026-07-05T08:39:33Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"DqdH1+hrs6sYalhFMRNksTW75vv46M4RxuMdLOpS+/lfjGVaEvgA4mrijJTLMAeAr1VIwxhe/21MbnmFENleDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T18:34:33.288943Z"},"content_sha256":"b96e74e684c359d6a7e6db2430bbdfdd6a0588eb9ac0507e05f6c2555ff1acb1","schema_version":"1.0","event_id":"sha256:b96e74e684c359d6a7e6db2430bbdfdd6a0588eb9ac0507e05f6c2555ff1acb1"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/26ZTRR2F22FANJVBUOTQVLTJMI/bundle.json","state_url":"https://pith.science/pith/26ZTRR2F22FANJVBUOTQVLTJMI/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/26ZTRR2F22FANJVBUOTQVLTJMI/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":"2026-08-09T18:34:33Z","links":{"resolver":"https://pith.science/pith/26ZTRR2F22FANJVBUOTQVLTJMI","bundle":"https://pith.science/pith/26ZTRR2F22FANJVBUOTQVLTJMI/bundle.json","state":"https://pith.science/pith/26ZTRR2F22FANJVBUOTQVLTJMI/state.json","well_known_bundle":"https://pith.science/.well-known/pith/26ZTRR2F22FANJVBUOTQVLTJMI/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:26ZTRR2F22FANJVBUOTQVLTJMI","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":"ea311be10a3a1c437b7236e3a493e6b61f64b51aa558ce9d62b128e35483f1c6","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-08T21:26:22Z","title_canon_sha256":"ea0b933138ca0480367ac48901ae492f3871422db40a5c9ac676aeee2cc33106"},"schema_version":"1.0","source":{"id":"2404.05880","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.05880","created_at":"2026-07-05T08:39:33Z"},{"alias_kind":"arxiv_version","alias_value":"2404.05880v2","created_at":"2026-07-05T08:39:33Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.05880","created_at":"2026-07-05T08:39:33Z"},{"alias_kind":"pith_short_12","alias_value":"26ZTRR2F22FA","created_at":"2026-07-05T08:39:33Z"},{"alias_kind":"pith_short_16","alias_value":"26ZTRR2F22FANJVB","created_at":"2026-07-05T08:39:33Z"},{"alias_kind":"pith_short_8","alias_value":"26ZTRR2F","created_at":"2026-07-05T08:39:33Z"}],"graph_snapshots":[{"event_id":"sha256:b96e74e684c359d6a7e6db2430bbdfdd6a0588eb9ac0507e05f6c2555ff1acb1","target":"graph","created_at":"2026-07-05T08:39:33Z","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/2404.05880/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Jailbreaking attacks can enable Large Language Models (LLMs) to bypass the safeguard and generate harmful content. Existing jailbreaking defense methods have failed to address the fundamental issue that harmful knowledge resides within the model, leading to potential jailbreak risks for LLMs. In this paper, we propose a novel defense method called Eraser, which mainly includes three goals: unlearning harmful knowledge, retaining general knowledge, and maintaining safety alignment. The intuition is that if an LLM forgets the specific knowledge required to answer a harmful question, it will no l","authors_text":"Cen Chen, Huiping Zhuang, Jianwei Wang, Weikai Lu, Zelin Chen, Zhengdong Lu, Ziqian Zeng","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-08T21:26:22Z","title":"Eraser: Jailbreaking Defense in Large Language Models via Unlearning Harmful Knowledge"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.05880","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:a78ca18a964a35f55204091870495daea21cc9a47a3dab1f114ed696286e3aa4","target":"record","created_at":"2026-07-05T08:39:33Z","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":"ea311be10a3a1c437b7236e3a493e6b61f64b51aa558ce9d62b128e35483f1c6","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-08T21:26:22Z","title_canon_sha256":"ea0b933138ca0480367ac48901ae492f3871422db40a5c9ac676aeee2cc33106"},"schema_version":"1.0","source":{"id":"2404.05880","kind":"arxiv","version":2}},"canonical_sha256":"d7b338c745d68a06a6a1a3a70aae69622808f9c2c73f6042c6469a6ed110767e","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"d7b338c745d68a06a6a1a3a70aae69622808f9c2c73f6042c6469a6ed110767e","first_computed_at":"2026-07-05T08:39:33.339945Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:39:33.339945Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"iaasHWwIJWEesviTKgm0eBB/6j4+AXCfRHKQ6P8EEsv12HvHkhAg/bddMNIFQDbEIPPfFktnrf6Y5epXVygLBw==","signature_status":"signed_v1","signed_at":"2026-07-05T08:39:33.340433Z","signed_message":"canonical_sha256_bytes"},"source_id":"2404.05880","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a78ca18a964a35f55204091870495daea21cc9a47a3dab1f114ed696286e3aa4","sha256:b96e74e684c359d6a7e6db2430bbdfdd6a0588eb9ac0507e05f6c2555ff1acb1"],"state_sha256":"11d6a99b8704fc62f5554ff93af3b2d4221b00dfb17185f7fc170bde21a33000"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"iCGXMdGBfOASeeniH8nChYfjCK1WNGCHdQCxxbkCp865JBmh5uUfmVqDkIKdW1UE4Zm6Sns3DohpD5agtssDAQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T18:34:33.294221Z","bundle_sha256":"fb1ab4e3c00a5bb135459ab12592eac2062f3758f0c7f7ac59d38f73a57f9890"}}