{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:EGLX34FCUMWYMWTDRWTS3ITFVZ","short_pith_number":"pith:EGLX34FC","canonical_record":{"source":{"id":"2405.18540","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-28T19:16:17Z","cross_cats_sorted":["cs.CR","cs.LG"],"title_canon_sha256":"593e5a7335985fae38678c90fc3c09982246292c010560667ec1d7b6f7c798cc","abstract_canon_sha256":"e2e70b0eecd10ebb7780a99c4453d414577ed65f4649cdb2665123ae57aae437"},"schema_version":"1.0"},"canonical_sha256":"21977df0a2a32d865a638da72da265ae72d3f23e2897f4771abb242142571b14","source":{"kind":"arxiv","id":"2405.18540","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.18540","created_at":"2026-07-05T10:21:19Z"},{"alias_kind":"arxiv_version","alias_value":"2405.18540v2","created_at":"2026-07-05T10:21:19Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.18540","created_at":"2026-07-05T10:21:19Z"},{"alias_kind":"pith_short_12","alias_value":"EGLX34FCUMWY","created_at":"2026-07-05T10:21:19Z"},{"alias_kind":"pith_short_16","alias_value":"EGLX34FCUMWYMWTD","created_at":"2026-07-05T10:21:19Z"},{"alias_kind":"pith_short_8","alias_value":"EGLX34FC","created_at":"2026-07-05T10:21:19Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:EGLX34FCUMWYMWTDRWTS3ITFVZ","target":"record","payload":{"canonical_record":{"source":{"id":"2405.18540","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-28T19:16:17Z","cross_cats_sorted":["cs.CR","cs.LG"],"title_canon_sha256":"593e5a7335985fae38678c90fc3c09982246292c010560667ec1d7b6f7c798cc","abstract_canon_sha256":"e2e70b0eecd10ebb7780a99c4453d414577ed65f4649cdb2665123ae57aae437"},"schema_version":"1.0"},"canonical_sha256":"21977df0a2a32d865a638da72da265ae72d3f23e2897f4771abb242142571b14","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:21:19.568620Z","signature_b64":"DJHXsyMYqY8QcQsrAibwa4eR0s1Y/9H82v4OMEGFpRp7qhO/kGRNg/KbbaxaMzwew/7ldvmxkOQROQj5DREMAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"21977df0a2a32d865a638da72da265ae72d3f23e2897f4771abb242142571b14","last_reissued_at":"2026-07-05T10:21:19.568097Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:21:19.568097Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.18540","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":"2026-07-05T10:21:19Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"9C17um60NwgQhf6fGih7VtwnZAOzSa8acEfXinXqKD1Q4AiauNY2plQ6pWImwMF3I6EVKmcm1sbY3KjXzgTbAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-13T22:47:02.838234Z"},"content_sha256":"a8062400c72802dd3b40e27967be35c05fec86aa50cbf971e993893c94412300","schema_version":"1.0","event_id":"sha256:a8062400c72802dd3b40e27967be35c05fec86aa50cbf971e993893c94412300"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:EGLX34FCUMWYMWTDRWTS3ITFVZ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Learning diverse attacks on large language models for robust red-teaming and safety tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR","cs.LG"],"primary_cat":"cs.CL","authors_text":"David Dobre, Gauthier Gidel, Juho Lee, Kenji Kawaguchi, Lynn Cherif, Minsu Kim, Moksh Jain, Nikolay Malkin, Seanie Lee, Sung Ju Hwang, Yoshua Bengio","submitted_at":"2024-05-28T19:16:17Z","abstract_excerpt":"Red-teaming, or identifying prompts that elicit harmful responses, is a critical step in ensuring the safe and responsible deployment of large language models (LLMs). Developing effective protection against many modes of attack prompts requires discovering diverse attacks. Automated red-teaming typically uses reinforcement learning to fine-tune an attacker language model to generate prompts that elicit undesirable responses from a target LLM, as measured, for example, by an auxiliary toxicity classifier. We show that even with explicit regularization to favor novelty and diversity, existing ap"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.18540","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/2405.18540/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":"2026-07-05T10:21:19Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4woOt68IbVA/EHODxHUuEDLnTu3rNmMu6/IS8is4S2SgsDDtgyCR3yU6QD8oee5a6j6Cu5W7e1d8SLTclrcKBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-13T22:47:02.838627Z"},"content_sha256":"1cf1f27fcee2364862eae7f27a94a1ff074e6d06f81e33250369bd946cbdb2b7","schema_version":"1.0","event_id":"sha256:1cf1f27fcee2364862eae7f27a94a1ff074e6d06f81e33250369bd946cbdb2b7"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/EGLX34FCUMWYMWTDRWTS3ITFVZ/bundle.json","state_url":"https://pith.science/pith/EGLX34FCUMWYMWTDRWTS3ITFVZ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/EGLX34FCUMWYMWTDRWTS3ITFVZ/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":"2026-07-13T22:47:02Z","links":{"resolver":"https://pith.science/pith/EGLX34FCUMWYMWTDRWTS3ITFVZ","bundle":"https://pith.science/pith/EGLX34FCUMWYMWTDRWTS3ITFVZ/bundle.json","state":"https://pith.science/pith/EGLX34FCUMWYMWTDRWTS3ITFVZ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/EGLX34FCUMWYMWTDRWTS3ITFVZ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:EGLX34FCUMWYMWTDRWTS3ITFVZ","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":"e2e70b0eecd10ebb7780a99c4453d414577ed65f4649cdb2665123ae57aae437","cross_cats_sorted":["cs.CR","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-28T19:16:17Z","title_canon_sha256":"593e5a7335985fae38678c90fc3c09982246292c010560667ec1d7b6f7c798cc"},"schema_version":"1.0","source":{"id":"2405.18540","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.18540","created_at":"2026-07-05T10:21:19Z"},{"alias_kind":"arxiv_version","alias_value":"2405.18540v2","created_at":"2026-07-05T10:21:19Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.18540","created_at":"2026-07-05T10:21:19Z"},{"alias_kind":"pith_short_12","alias_value":"EGLX34FCUMWY","created_at":"2026-07-05T10:21:19Z"},{"alias_kind":"pith_short_16","alias_value":"EGLX34FCUMWYMWTD","created_at":"2026-07-05T10:21:19Z"},{"alias_kind":"pith_short_8","alias_value":"EGLX34FC","created_at":"2026-07-05T10:21:19Z"}],"graph_snapshots":[{"event_id":"sha256:1cf1f27fcee2364862eae7f27a94a1ff074e6d06f81e33250369bd946cbdb2b7","target":"graph","created_at":"2026-07-05T10:21:19Z","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/2405.18540/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Red-teaming, or identifying prompts that elicit harmful responses, is a critical step in ensuring the safe and responsible deployment of large language models (LLMs). Developing effective protection against many modes of attack prompts requires discovering diverse attacks. Automated red-teaming typically uses reinforcement learning to fine-tune an attacker language model to generate prompts that elicit undesirable responses from a target LLM, as measured, for example, by an auxiliary toxicity classifier. We show that even with explicit regularization to favor novelty and diversity, existing ap","authors_text":"David Dobre, Gauthier Gidel, Juho Lee, Kenji Kawaguchi, Lynn Cherif, Minsu Kim, Moksh Jain, Nikolay Malkin, Seanie Lee, Sung Ju Hwang, Yoshua Bengio","cross_cats":["cs.CR","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-28T19:16:17Z","title":"Learning diverse attacks on large language models for robust red-teaming and safety tuning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.18540","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:a8062400c72802dd3b40e27967be35c05fec86aa50cbf971e993893c94412300","target":"record","created_at":"2026-07-05T10:21:19Z","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":"e2e70b0eecd10ebb7780a99c4453d414577ed65f4649cdb2665123ae57aae437","cross_cats_sorted":["cs.CR","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-28T19:16:17Z","title_canon_sha256":"593e5a7335985fae38678c90fc3c09982246292c010560667ec1d7b6f7c798cc"},"schema_version":"1.0","source":{"id":"2405.18540","kind":"arxiv","version":2}},"canonical_sha256":"21977df0a2a32d865a638da72da265ae72d3f23e2897f4771abb242142571b14","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"21977df0a2a32d865a638da72da265ae72d3f23e2897f4771abb242142571b14","first_computed_at":"2026-07-05T10:21:19.568097Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:21:19.568097Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"DJHXsyMYqY8QcQsrAibwa4eR0s1Y/9H82v4OMEGFpRp7qhO/kGRNg/KbbaxaMzwew/7ldvmxkOQROQj5DREMAA==","signature_status":"signed_v1","signed_at":"2026-07-05T10:21:19.568620Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.18540","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a8062400c72802dd3b40e27967be35c05fec86aa50cbf971e993893c94412300","sha256:1cf1f27fcee2364862eae7f27a94a1ff074e6d06f81e33250369bd946cbdb2b7"],"state_sha256":"42e55f43474f10c6fbd417dce944f34a5fda6ce911fd643f072937f9490ae66f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"TadBS1OWMnVQtNNjerb3HV4AB4b7eAlSEcdE1BWXMDkeDQyTYgdh8M/qm922YkTDGK+BwFFBXKJ4xlxm3544Cw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-13T22:47:02.840815Z","bundle_sha256":"19349ccb123abbc77235e37941e2124342892b830a5778b646f7249a44bb9cf5"}}