{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:RJHXPCUZXXY2AYXQC4ZND7ZBV6","short_pith_number":"pith:RJHXPCUZ","canonical_record":{"source":{"id":"2405.19026","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-29T12:12:09Z","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"title_canon_sha256":"1103022eec654168c584bdfb6691d7af882349e8ffbb2fe2f80483b8d8de737d","abstract_canon_sha256":"ce4664b3faabf4286676dd6552c78deee7459c3976213f1f8279d8a4fbf3306a"},"schema_version":"1.0"},"canonical_sha256":"8a4f778a99bdf1a062f01732d1ff21afb735a7c27846adb085da040665f4f6f5","source":{"kind":"arxiv","id":"2405.19026","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.19026","created_at":"2026-07-05T09:52:11Z"},{"alias_kind":"arxiv_version","alias_value":"2405.19026v2","created_at":"2026-07-05T09:52:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.19026","created_at":"2026-07-05T09:52:11Z"},{"alias_kind":"pith_short_12","alias_value":"RJHXPCUZXXY2","created_at":"2026-07-05T09:52:11Z"},{"alias_kind":"pith_short_16","alias_value":"RJHXPCUZXXY2AYXQ","created_at":"2026-07-05T09:52:11Z"},{"alias_kind":"pith_short_8","alias_value":"RJHXPCUZ","created_at":"2026-07-05T09:52:11Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:RJHXPCUZXXY2AYXQC4ZND7ZBV6","target":"record","payload":{"canonical_record":{"source":{"id":"2405.19026","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-29T12:12:09Z","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"title_canon_sha256":"1103022eec654168c584bdfb6691d7af882349e8ffbb2fe2f80483b8d8de737d","abstract_canon_sha256":"ce4664b3faabf4286676dd6552c78deee7459c3976213f1f8279d8a4fbf3306a"},"schema_version":"1.0"},"canonical_sha256":"8a4f778a99bdf1a062f01732d1ff21afb735a7c27846adb085da040665f4f6f5","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:52:11.989933Z","signature_b64":"BYiXWXs3+gj+dC6YpyGUEicKe/dVXH6lC9QLBAWPHe3rMcFR+oclBxhAGewnPKiJN2hR0lhG1naFsIk8djYPDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8a4f778a99bdf1a062f01732d1ff21afb735a7c27846adb085da040665f4f6f5","last_reissued_at":"2026-07-05T09:52:11.989506Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:52:11.989506Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.19026","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":"2026-07-05T09:52:11Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Yab6kEs6RWYD3F09HIEyXcB/a13AqKDeLcKUJmmcnPjIF0j9v5gRST6osp7rCYxJ/O2D/CcAbqoLai40FdBTDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T23:29:49.874268Z"},"content_sha256":"9767699bba24366f405a4af3b8b1a10a8736525f7476a68e77de9fafb83728b4","schema_version":"1.0","event_id":"sha256:9767699bba24366f405a4af3b8b1a10a8736525f7476a68e77de9fafb83728b4"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:RJHXPCUZXXY2AYXQC4ZND7ZBV6","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"DiveR-CT: Diversity-enhanced Red Teaming Large Language Model Assistants with Relaxing Constraints","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.CR"],"primary_cat":"cs.LG","authors_text":"Andrew Zhao, Gao Huang, Matthieu Lin, Quentin Xu, Shenzhi Wang, Yong-jin Liu, Zilong Zheng","submitted_at":"2024-05-29T12:12:09Z","abstract_excerpt":"Recent advances in large language model assistants have made them indispensable, raising significant concerns over managing their safety. Automated red teaming offers a promising alternative to the labor-intensive and error-prone manual probing for vulnerabilities, providing more consistent and scalable safety evaluations. However, existing approaches often compromise diversity by focusing on maximizing attack success rate. Additionally, methods that decrease the cosine similarity from historical embeddings with semantic diversity rewards lead to novelty stagnation as history grows. To address"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.19026","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/2405.19026/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":"2026-07-05T09:52:11Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"w2kQXCLnMJ9EXDZQMbInO134RrbPI+oeic4UZs23m13p56uzyP+Np3iChqxOBKqagq6ytxLZQovQOd7lSerXDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T23:29:49.875155Z"},"content_sha256":"1cac49b722fae70017120470d75a19d47eeb3cc5dea87b7e8802f734c0b2e952","schema_version":"1.0","event_id":"sha256:1cac49b722fae70017120470d75a19d47eeb3cc5dea87b7e8802f734c0b2e952"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/RJHXPCUZXXY2AYXQC4ZND7ZBV6/bundle.json","state_url":"https://pith.science/pith/RJHXPCUZXXY2AYXQC4ZND7ZBV6/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/RJHXPCUZXXY2AYXQC4ZND7ZBV6/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":"2026-08-10T23:29:49Z","links":{"resolver":"https://pith.science/pith/RJHXPCUZXXY2AYXQC4ZND7ZBV6","bundle":"https://pith.science/pith/RJHXPCUZXXY2AYXQC4ZND7ZBV6/bundle.json","state":"https://pith.science/pith/RJHXPCUZXXY2AYXQC4ZND7ZBV6/state.json","well_known_bundle":"https://pith.science/.well-known/pith/RJHXPCUZXXY2AYXQC4ZND7ZBV6/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:RJHXPCUZXXY2AYXQC4ZND7ZBV6","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":"ce4664b3faabf4286676dd6552c78deee7459c3976213f1f8279d8a4fbf3306a","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-29T12:12:09Z","title_canon_sha256":"1103022eec654168c584bdfb6691d7af882349e8ffbb2fe2f80483b8d8de737d"},"schema_version":"1.0","source":{"id":"2405.19026","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.19026","created_at":"2026-07-05T09:52:11Z"},{"alias_kind":"arxiv_version","alias_value":"2405.19026v2","created_at":"2026-07-05T09:52:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.19026","created_at":"2026-07-05T09:52:11Z"},{"alias_kind":"pith_short_12","alias_value":"RJHXPCUZXXY2","created_at":"2026-07-05T09:52:11Z"},{"alias_kind":"pith_short_16","alias_value":"RJHXPCUZXXY2AYXQ","created_at":"2026-07-05T09:52:11Z"},{"alias_kind":"pith_short_8","alias_value":"RJHXPCUZ","created_at":"2026-07-05T09:52:11Z"}],"graph_snapshots":[{"event_id":"sha256:1cac49b722fae70017120470d75a19d47eeb3cc5dea87b7e8802f734c0b2e952","target":"graph","created_at":"2026-07-05T09:52:11Z","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/2405.19026/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent advances in large language model assistants have made them indispensable, raising significant concerns over managing their safety. Automated red teaming offers a promising alternative to the labor-intensive and error-prone manual probing for vulnerabilities, providing more consistent and scalable safety evaluations. However, existing approaches often compromise diversity by focusing on maximizing attack success rate. Additionally, methods that decrease the cosine similarity from historical embeddings with semantic diversity rewards lead to novelty stagnation as history grows. To address","authors_text":"Andrew Zhao, Gao Huang, Matthieu Lin, Quentin Xu, Shenzhi Wang, Yong-jin Liu, Zilong Zheng","cross_cats":["cs.AI","cs.CL","cs.CR"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-29T12:12:09Z","title":"DiveR-CT: Diversity-enhanced Red Teaming Large Language Model Assistants with Relaxing Constraints"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.19026","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":"sha256:9767699bba24366f405a4af3b8b1a10a8736525f7476a68e77de9fafb83728b4","target":"record","created_at":"2026-07-05T09:52:11Z","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":"ce4664b3faabf4286676dd6552c78deee7459c3976213f1f8279d8a4fbf3306a","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-29T12:12:09Z","title_canon_sha256":"1103022eec654168c584bdfb6691d7af882349e8ffbb2fe2f80483b8d8de737d"},"schema_version":"1.0","source":{"id":"2405.19026","kind":"arxiv","version":2}},"canonical_sha256":"8a4f778a99bdf1a062f01732d1ff21afb735a7c27846adb085da040665f4f6f5","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"8a4f778a99bdf1a062f01732d1ff21afb735a7c27846adb085da040665f4f6f5","first_computed_at":"2026-07-05T09:52:11.989506Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:52:11.989506Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"BYiXWXs3+gj+dC6YpyGUEicKe/dVXH6lC9QLBAWPHe3rMcFR+oclBxhAGewnPKiJN2hR0lhG1naFsIk8djYPDQ==","signature_status":"signed_v1","signed_at":"2026-07-05T09:52:11.989933Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.19026","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9767699bba24366f405a4af3b8b1a10a8736525f7476a68e77de9fafb83728b4","sha256:1cac49b722fae70017120470d75a19d47eeb3cc5dea87b7e8802f734c0b2e952"],"state_sha256":"593cf56e580ba8e8b98fe1169d83f79afb12b5b020a2af8f87fb0c9ae8544487"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"6Rm43quuu/4EIay/6RyTE0sQEZrx7nWo/AeXRIWpXZ4+W2GF2p8+2WozAraI8c1g8Co/aQs+gZydAlJUz1mKDA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-10T23:29:49.879254Z","bundle_sha256":"803575b9438c9e3a7f876e675196b74bb3b20bd944edb3b23b2ebfe943b286cc"}}