{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:IE3I4HUQ33N6AXLYUHLTFD3LL5","short_pith_number":"pith:IE3I4HUQ","canonical_record":{"source":{"id":"2403.04783","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-03-02T16:52:22Z","cross_cats_sorted":["cs.CL","cs.CR"],"title_canon_sha256":"da96d90fe66a607e18c59cb055307906bb59d01b9813c2990913344409cb3ee6","abstract_canon_sha256":"d7dfd4ab33ebc65dff5095319aea8ce52942134d567ec8e3cd635623a838c1bd"},"schema_version":"1.0"},"canonical_sha256":"41368e1e90dedbe05d78a1d7328f6b5f56a4f57af56711578e0a382efeb69c13","source":{"kind":"arxiv","id":"2403.04783","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2403.04783","created_at":"2026-07-05T09:35:08Z"},{"alias_kind":"arxiv_version","alias_value":"2403.04783v2","created_at":"2026-07-05T09:35:08Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.04783","created_at":"2026-07-05T09:35:08Z"},{"alias_kind":"pith_short_12","alias_value":"IE3I4HUQ33N6","created_at":"2026-07-05T09:35:08Z"},{"alias_kind":"pith_short_16","alias_value":"IE3I4HUQ33N6AXLY","created_at":"2026-07-05T09:35:08Z"},{"alias_kind":"pith_short_8","alias_value":"IE3I4HUQ","created_at":"2026-07-05T09:35:08Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:IE3I4HUQ33N6AXLYUHLTFD3LL5","target":"record","payload":{"canonical_record":{"source":{"id":"2403.04783","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-03-02T16:52:22Z","cross_cats_sorted":["cs.CL","cs.CR"],"title_canon_sha256":"da96d90fe66a607e18c59cb055307906bb59d01b9813c2990913344409cb3ee6","abstract_canon_sha256":"d7dfd4ab33ebc65dff5095319aea8ce52942134d567ec8e3cd635623a838c1bd"},"schema_version":"1.0"},"canonical_sha256":"41368e1e90dedbe05d78a1d7328f6b5f56a4f57af56711578e0a382efeb69c13","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:35:08.702751Z","signature_b64":"JHG4mvT0JqkLye41n1L6yDwin/SUWe8qYuKzXG7I2HCzDt0Dm/fAkcX4wyyArgvtaWuAfZ8l6rGPbEPL71osDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"41368e1e90dedbe05d78a1d7328f6b5f56a4f57af56711578e0a382efeb69c13","last_reissued_at":"2026-07-05T09:35:08.702300Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:35:08.702300Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2403.04783","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":"2026-07-05T09:35:08Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"VIeJtELm79M/xbObtKRCMQR4Hh4+otSWAehUhkTScWbyNQn8BWEJ82FHpNlI4NLHtxfoDy4/kpaHrTNcFoz1DQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T10:40:35.938247Z"},"content_sha256":"0e5a389465b000c038a3042adec9e59ee0e01e74864eb93921f2b6b3a16be3f2","schema_version":"1.0","event_id":"sha256:0e5a389465b000c038a3042adec9e59ee0e01e74864eb93921f2b6b3a16be3f2"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:IE3I4HUQ33N6AXLYUHLTFD3LL5","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"AutoDefense: Multi-Agent LLM Defense against Jailbreak Attacks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.CR"],"primary_cat":"cs.LG","authors_text":"Huazheng Wang, Qingyun Wu, Xiao Zhang, Yifan Zeng, Yiran Wu","submitted_at":"2024-03-02T16:52:22Z","abstract_excerpt":"Despite extensive pre-training in moral alignment to prevent generating harmful information, large language models (LLMs) remain vulnerable to jailbreak attacks. In this paper, we propose AutoDefense, a multi-agent defense framework that filters harmful responses from LLMs. With the response-filtering mechanism, our framework is robust against different jailbreak attack prompts, and can be used to defend different victim models. AutoDefense assigns different roles to LLM agents and employs them to complete the defense task collaboratively. The division in tasks enhances the overall instruction"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.04783","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/2403.04783/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":"2026-07-05T09:35:08Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"oN6qsmBWIJZ/PH866b8/DWvyy++eha5PbVxovC5H2bkTTurYskW32Dldkn4RQm9tk/DSk+TrLhk7P6x6DjbHCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T10:40:35.938788Z"},"content_sha256":"8d2380d7525269d71e70ec1d7147abc06fdebdf7ecc52804453e462fffa8d11e","schema_version":"1.0","event_id":"sha256:8d2380d7525269d71e70ec1d7147abc06fdebdf7ecc52804453e462fffa8d11e"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/IE3I4HUQ33N6AXLYUHLTFD3LL5/bundle.json","state_url":"https://pith.science/pith/IE3I4HUQ33N6AXLYUHLTFD3LL5/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/IE3I4HUQ33N6AXLYUHLTFD3LL5/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":"2026-08-05T10:40:35Z","links":{"resolver":"https://pith.science/pith/IE3I4HUQ33N6AXLYUHLTFD3LL5","bundle":"https://pith.science/pith/IE3I4HUQ33N6AXLYUHLTFD3LL5/bundle.json","state":"https://pith.science/pith/IE3I4HUQ33N6AXLYUHLTFD3LL5/state.json","well_known_bundle":"https://pith.science/.well-known/pith/IE3I4HUQ33N6AXLYUHLTFD3LL5/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:IE3I4HUQ33N6AXLYUHLTFD3LL5","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":"d7dfd4ab33ebc65dff5095319aea8ce52942134d567ec8e3cd635623a838c1bd","cross_cats_sorted":["cs.CL","cs.CR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-03-02T16:52:22Z","title_canon_sha256":"da96d90fe66a607e18c59cb055307906bb59d01b9813c2990913344409cb3ee6"},"schema_version":"1.0","source":{"id":"2403.04783","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2403.04783","created_at":"2026-07-05T09:35:08Z"},{"alias_kind":"arxiv_version","alias_value":"2403.04783v2","created_at":"2026-07-05T09:35:08Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.04783","created_at":"2026-07-05T09:35:08Z"},{"alias_kind":"pith_short_12","alias_value":"IE3I4HUQ33N6","created_at":"2026-07-05T09:35:08Z"},{"alias_kind":"pith_short_16","alias_value":"IE3I4HUQ33N6AXLY","created_at":"2026-07-05T09:35:08Z"},{"alias_kind":"pith_short_8","alias_value":"IE3I4HUQ","created_at":"2026-07-05T09:35:08Z"}],"graph_snapshots":[{"event_id":"sha256:8d2380d7525269d71e70ec1d7147abc06fdebdf7ecc52804453e462fffa8d11e","target":"graph","created_at":"2026-07-05T09:35:08Z","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/2403.04783/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Despite extensive pre-training in moral alignment to prevent generating harmful information, large language models (LLMs) remain vulnerable to jailbreak attacks. In this paper, we propose AutoDefense, a multi-agent defense framework that filters harmful responses from LLMs. With the response-filtering mechanism, our framework is robust against different jailbreak attack prompts, and can be used to defend different victim models. AutoDefense assigns different roles to LLM agents and employs them to complete the defense task collaboratively. The division in tasks enhances the overall instruction","authors_text":"Huazheng Wang, Qingyun Wu, Xiao Zhang, Yifan Zeng, Yiran Wu","cross_cats":["cs.CL","cs.CR"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-03-02T16:52:22Z","title":"AutoDefense: Multi-Agent LLM Defense against Jailbreak Attacks"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.04783","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":"sha256:0e5a389465b000c038a3042adec9e59ee0e01e74864eb93921f2b6b3a16be3f2","target":"record","created_at":"2026-07-05T09:35:08Z","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":"d7dfd4ab33ebc65dff5095319aea8ce52942134d567ec8e3cd635623a838c1bd","cross_cats_sorted":["cs.CL","cs.CR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-03-02T16:52:22Z","title_canon_sha256":"da96d90fe66a607e18c59cb055307906bb59d01b9813c2990913344409cb3ee6"},"schema_version":"1.0","source":{"id":"2403.04783","kind":"arxiv","version":2}},"canonical_sha256":"41368e1e90dedbe05d78a1d7328f6b5f56a4f57af56711578e0a382efeb69c13","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"41368e1e90dedbe05d78a1d7328f6b5f56a4f57af56711578e0a382efeb69c13","first_computed_at":"2026-07-05T09:35:08.702300Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:35:08.702300Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"JHG4mvT0JqkLye41n1L6yDwin/SUWe8qYuKzXG7I2HCzDt0Dm/fAkcX4wyyArgvtaWuAfZ8l6rGPbEPL71osDQ==","signature_status":"signed_v1","signed_at":"2026-07-05T09:35:08.702751Z","signed_message":"canonical_sha256_bytes"},"source_id":"2403.04783","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0e5a389465b000c038a3042adec9e59ee0e01e74864eb93921f2b6b3a16be3f2","sha256:8d2380d7525269d71e70ec1d7147abc06fdebdf7ecc52804453e462fffa8d11e"],"state_sha256":"82924fe5017ace96ce6fc3a5de6a63bf5606ce4b3c8bbb74e63e34fb1075bcf7"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"J/gyWBg2QI/coVCUjfIzHP21BBPVvtqb1APP1U/FzpAzzq5f+jcIdQ+uKSF1Kl87nROF+CGOz3DeMPJfjaJxDA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-05T10:40:35.942886Z","bundle_sha256":"5cd41bd41cab6b5153b27864f228a7a11e17ac3b3db6eaeb79b672ff04217c68"}}