{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:BUQUHOYQPOI6KUHDSFFNZW5RMF","short_pith_number":"pith:BUQUHOYQ","canonical_record":{"source":{"id":"2310.15140","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2023-10-23T17:46:07Z","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"title_canon_sha256":"d5b5989ed1e3ef5af4fd1d1e4b91d152f802aeb9b0ece8db9fefcc99fed9c59c","abstract_canon_sha256":"2cdd0660681f0f8b2a90adcb75086f7ad0d7c3e76cf3518c487d700a7cbee5ba"},"schema_version":"1.0"},"canonical_sha256":"0d2143bb107b91e550e3914adcdbb16162b5ded1b051e53ea6cc35bc47c4594c","source":{"kind":"arxiv","id":"2310.15140","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2310.15140","created_at":"2026-07-05T07:24:01Z"},{"alias_kind":"arxiv_version","alias_value":"2310.15140v2","created_at":"2026-07-05T07:24:01Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.15140","created_at":"2026-07-05T07:24:01Z"},{"alias_kind":"pith_short_12","alias_value":"BUQUHOYQPOI6","created_at":"2026-07-05T07:24:01Z"},{"alias_kind":"pith_short_16","alias_value":"BUQUHOYQPOI6KUHD","created_at":"2026-07-05T07:24:01Z"},{"alias_kind":"pith_short_8","alias_value":"BUQUHOYQ","created_at":"2026-07-05T07:24:01Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:BUQUHOYQPOI6KUHDSFFNZW5RMF","target":"record","payload":{"canonical_record":{"source":{"id":"2310.15140","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2023-10-23T17:46:07Z","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"title_canon_sha256":"d5b5989ed1e3ef5af4fd1d1e4b91d152f802aeb9b0ece8db9fefcc99fed9c59c","abstract_canon_sha256":"2cdd0660681f0f8b2a90adcb75086f7ad0d7c3e76cf3518c487d700a7cbee5ba"},"schema_version":"1.0"},"canonical_sha256":"0d2143bb107b91e550e3914adcdbb16162b5ded1b051e53ea6cc35bc47c4594c","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:24:01.369099Z","signature_b64":"DkS9k/GcNPU2+VxxSNMezErfObUohFqYrRAB3CRM54LuEjFVnUDDnh8tiTSdAPRnoesCNa/YQZYI9QedyFLQCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0d2143bb107b91e550e3914adcdbb16162b5ded1b051e53ea6cc35bc47c4594c","last_reissued_at":"2026-07-05T07:24:01.368566Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:24:01.368566Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2310.15140","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":"2026-07-05T07:24:01Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"N897T+MZuBsW0vI/e/WPMoWmn/RCwUPJCtiU5pwnvIQvweKypZfmAiBmczM3Y650L7UR0gym2IGspSwIeG9aBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T01:29:59.499462Z"},"content_sha256":"ef1405080ae3d4fe0465b705a531ab5a3116a46cf3997389c501e156755052ba","schema_version":"1.0","event_id":"sha256:ef1405080ae3d4fe0465b705a531ab5a3116a46cf3997389c501e156755052ba"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:BUQUHOYQPOI6KUHDSFFNZW5RMF","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"AutoDAN: Interpretable Gradient-Based Adversarial Attacks on Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.CR","authors_text":"Ani Nenkova, Bang An, Furong Huang, Gang Wu, Joe Barrow, Ruiyi Zhang, Sicheng Zhu, Tong Sun, Zichao Wang","submitted_at":"2023-10-23T17:46:07Z","abstract_excerpt":"Safety alignment of Large Language Models (LLMs) can be compromised with manual jailbreak attacks and (automatic) adversarial attacks. Recent studies suggest that defending against these attacks is possible: adversarial attacks generate unlimited but unreadable gibberish prompts, detectable by perplexity-based filters; manual jailbreak attacks craft readable prompts, but their limited number due to the necessity of human creativity allows for easy blocking. In this paper, we show that these solutions may be too optimistic. We introduce AutoDAN, an interpretable, gradient-based adversarial atta"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.15140","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/2310.15140/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":"2026-07-05T07:24:01Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"eH8eTsLdf8lutTDpVqPv8UfdcCbQ125LZjgJkUQpSLjHo1vymLmR9ms/bP93LOQDVp1724iTJeBsq3mMBy7FCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T01:29:59.499964Z"},"content_sha256":"9319231bdd7d3af9d68639e68a8cc72a324b729b4e051916c7e07a7eff9caa44","schema_version":"1.0","event_id":"sha256:9319231bdd7d3af9d68639e68a8cc72a324b729b4e051916c7e07a7eff9caa44"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/BUQUHOYQPOI6KUHDSFFNZW5RMF/bundle.json","state_url":"https://pith.science/pith/BUQUHOYQPOI6KUHDSFFNZW5RMF/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/BUQUHOYQPOI6KUHDSFFNZW5RMF/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":"2026-08-04T01:29:59Z","links":{"resolver":"https://pith.science/pith/BUQUHOYQPOI6KUHDSFFNZW5RMF","bundle":"https://pith.science/pith/BUQUHOYQPOI6KUHDSFFNZW5RMF/bundle.json","state":"https://pith.science/pith/BUQUHOYQPOI6KUHDSFFNZW5RMF/state.json","well_known_bundle":"https://pith.science/.well-known/pith/BUQUHOYQPOI6KUHDSFFNZW5RMF/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:BUQUHOYQPOI6KUHDSFFNZW5RMF","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":"2cdd0660681f0f8b2a90adcb75086f7ad0d7c3e76cf3518c487d700a7cbee5ba","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2023-10-23T17:46:07Z","title_canon_sha256":"d5b5989ed1e3ef5af4fd1d1e4b91d152f802aeb9b0ece8db9fefcc99fed9c59c"},"schema_version":"1.0","source":{"id":"2310.15140","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2310.15140","created_at":"2026-07-05T07:24:01Z"},{"alias_kind":"arxiv_version","alias_value":"2310.15140v2","created_at":"2026-07-05T07:24:01Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.15140","created_at":"2026-07-05T07:24:01Z"},{"alias_kind":"pith_short_12","alias_value":"BUQUHOYQPOI6","created_at":"2026-07-05T07:24:01Z"},{"alias_kind":"pith_short_16","alias_value":"BUQUHOYQPOI6KUHD","created_at":"2026-07-05T07:24:01Z"},{"alias_kind":"pith_short_8","alias_value":"BUQUHOYQ","created_at":"2026-07-05T07:24:01Z"}],"graph_snapshots":[{"event_id":"sha256:9319231bdd7d3af9d68639e68a8cc72a324b729b4e051916c7e07a7eff9caa44","target":"graph","created_at":"2026-07-05T07:24:01Z","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/2310.15140/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) can be compromised with manual jailbreak attacks and (automatic) adversarial attacks. Recent studies suggest that defending against these attacks is possible: adversarial attacks generate unlimited but unreadable gibberish prompts, detectable by perplexity-based filters; manual jailbreak attacks craft readable prompts, but their limited number due to the necessity of human creativity allows for easy blocking. In this paper, we show that these solutions may be too optimistic. We introduce AutoDAN, an interpretable, gradient-based adversarial atta","authors_text":"Ani Nenkova, Bang An, Furong Huang, Gang Wu, Joe Barrow, Ruiyi Zhang, Sicheng Zhu, Tong Sun, Zichao Wang","cross_cats":["cs.AI","cs.CL","cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2023-10-23T17:46:07Z","title":"AutoDAN: Interpretable Gradient-Based Adversarial Attacks on Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.15140","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:ef1405080ae3d4fe0465b705a531ab5a3116a46cf3997389c501e156755052ba","target":"record","created_at":"2026-07-05T07:24:01Z","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":"2cdd0660681f0f8b2a90adcb75086f7ad0d7c3e76cf3518c487d700a7cbee5ba","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2023-10-23T17:46:07Z","title_canon_sha256":"d5b5989ed1e3ef5af4fd1d1e4b91d152f802aeb9b0ece8db9fefcc99fed9c59c"},"schema_version":"1.0","source":{"id":"2310.15140","kind":"arxiv","version":2}},"canonical_sha256":"0d2143bb107b91e550e3914adcdbb16162b5ded1b051e53ea6cc35bc47c4594c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0d2143bb107b91e550e3914adcdbb16162b5ded1b051e53ea6cc35bc47c4594c","first_computed_at":"2026-07-05T07:24:01.368566Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:24:01.368566Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"DkS9k/GcNPU2+VxxSNMezErfObUohFqYrRAB3CRM54LuEjFVnUDDnh8tiTSdAPRnoesCNa/YQZYI9QedyFLQCw==","signature_status":"signed_v1","signed_at":"2026-07-05T07:24:01.369099Z","signed_message":"canonical_sha256_bytes"},"source_id":"2310.15140","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:ef1405080ae3d4fe0465b705a531ab5a3116a46cf3997389c501e156755052ba","sha256:9319231bdd7d3af9d68639e68a8cc72a324b729b4e051916c7e07a7eff9caa44"],"state_sha256":"e2681362ac96e7fbe2caf1b0e607f1e726975c0fb49ec089f1f9849d30686b8f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"x5E8KoSq/U64vShfnUKXYaDOGmFras2WjH2dkmPGXFokghXCd3bSX2MT9HJBx5lITP9B5cpQOp8f85XJF1u1Ag==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T01:29:59.503340Z","bundle_sha256":"b55f9114d1f9a2c9ca639773446c3a8067d5f49f18a3662d8c163c0874d17228"}}