{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:ZNSOJVBLENAKSPY5KNRVBNURM7","short_pith_number":"pith:ZNSOJVBL","canonical_record":{"source":{"id":"2506.04302","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-06-04T14:50:24Z","cross_cats_sorted":[],"title_canon_sha256":"021dce96743b3204a00e6c691d51f1e4f972935b99a92b8d78bbca861e4297ff","abstract_canon_sha256":"2320591b3c37af2b463275423cf074d9257875b055672f0e02e705552fccf002"},"schema_version":"1.0"},"canonical_sha256":"cb64e4d42b2340a93f1d536350b69167f3419e6812abe397ac50188ad89bf871","source":{"kind":"arxiv","id":"2506.04302","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.04302","created_at":"2026-07-05T11:16:12Z"},{"alias_kind":"arxiv_version","alias_value":"2506.04302v1","created_at":"2026-07-05T11:16:12Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.04302","created_at":"2026-07-05T11:16:12Z"},{"alias_kind":"pith_short_12","alias_value":"ZNSOJVBLENAK","created_at":"2026-07-05T11:16:12Z"},{"alias_kind":"pith_short_16","alias_value":"ZNSOJVBLENAKSPY5","created_at":"2026-07-05T11:16:12Z"},{"alias_kind":"pith_short_8","alias_value":"ZNSOJVBL","created_at":"2026-07-05T11:16:12Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:ZNSOJVBLENAKSPY5KNRVBNURM7","target":"record","payload":{"canonical_record":{"source":{"id":"2506.04302","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-06-04T14:50:24Z","cross_cats_sorted":[],"title_canon_sha256":"021dce96743b3204a00e6c691d51f1e4f972935b99a92b8d78bbca861e4297ff","abstract_canon_sha256":"2320591b3c37af2b463275423cf074d9257875b055672f0e02e705552fccf002"},"schema_version":"1.0"},"canonical_sha256":"cb64e4d42b2340a93f1d536350b69167f3419e6812abe397ac50188ad89bf871","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:16:12.052371Z","signature_b64":"hqZQiHTGuNoPC9d//JSF5Z7SxBFUmW97GkXuw3YBTzufW8/K1Yee8rN7R/TIZ51773UjVGnucSKYdTzxWhVpAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cb64e4d42b2340a93f1d536350b69167f3419e6812abe397ac50188ad89bf871","last_reissued_at":"2026-07-05T11:16:12.051864Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:16:12.051864Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2506.04302","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:16:12Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Z+ijzCgsXS40eRjbqhsX8ezXOqCZBkSHeFSKQZ9UwBhAAGKiGXLvlHIcdyVllIxpbggSqHu8gSKAWxYi6N7SCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-19T22:02:37.852969Z"},"content_sha256":"4584fc0e7454fe7be6379b0d19c7b3c33b5d87dc6481935efebd6ada1dab5683","schema_version":"1.0","event_id":"sha256:4584fc0e7454fe7be6379b0d19c7b3c33b5d87dc6481935efebd6ada1dab5683"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:ZNSOJVBLENAKSPY5KNRVBNURM7","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"RedRFT: A Light-Weight Benchmark for Reinforcement Fine-Tuning-Based Red Teaming","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Cong Wang, Wei-Bin Lee, Xiang Zheng, Xingjun Ma","submitted_at":"2025-06-04T14:50:24Z","abstract_excerpt":"Red teaming has proven to be an effective method for identifying and mitigating vulnerabilities in Large Language Models (LLMs). Reinforcement Fine-Tuning (RFT) has emerged as a promising strategy among existing red teaming techniques. However, a lack of a unified benchmark hinders current RFT-based red teaming methods. Implementation details, especially in Proximal Policy Optimization (PPO)-based RFT, significantly affect outcome stability and reproducibility. To address this issue, we introduce RedRFT, a lightweight benchmark designed to simplify and standardize the implementation and evalua"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.04302","kind":"arxiv","version":1},"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/2506.04302/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":"2026-07-05T11:16:12Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+4J8og5urc6v1YDuLfvgdHqMDQ10h9281kD/vOPXl52rrJWNPN0+Ox44uQ8YRyI4Pfh09Qr82ZnrLrWESWHEBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-19T22:02:37.853601Z"},"content_sha256":"3db16c7b831b861c8221cfe6cacfb0f2a3b8851e3b5e5d095c7b9c076b000733","schema_version":"1.0","event_id":"sha256:3db16c7b831b861c8221cfe6cacfb0f2a3b8851e3b5e5d095c7b9c076b000733"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/ZNSOJVBLENAKSPY5KNRVBNURM7/bundle.json","state_url":"https://pith.science/pith/ZNSOJVBLENAKSPY5KNRVBNURM7/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/ZNSOJVBLENAKSPY5KNRVBNURM7/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":"2026-08-19T22:02:37Z","links":{"resolver":"https://pith.science/pith/ZNSOJVBLENAKSPY5KNRVBNURM7","bundle":"https://pith.science/pith/ZNSOJVBLENAKSPY5KNRVBNURM7/bundle.json","state":"https://pith.science/pith/ZNSOJVBLENAKSPY5KNRVBNURM7/state.json","well_known_bundle":"https://pith.science/.well-known/pith/ZNSOJVBLENAKSPY5KNRVBNURM7/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:ZNSOJVBLENAKSPY5KNRVBNURM7","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":"2320591b3c37af2b463275423cf074d9257875b055672f0e02e705552fccf002","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-06-04T14:50:24Z","title_canon_sha256":"021dce96743b3204a00e6c691d51f1e4f972935b99a92b8d78bbca861e4297ff"},"schema_version":"1.0","source":{"id":"2506.04302","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.04302","created_at":"2026-07-05T11:16:12Z"},{"alias_kind":"arxiv_version","alias_value":"2506.04302v1","created_at":"2026-07-05T11:16:12Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.04302","created_at":"2026-07-05T11:16:12Z"},{"alias_kind":"pith_short_12","alias_value":"ZNSOJVBLENAK","created_at":"2026-07-05T11:16:12Z"},{"alias_kind":"pith_short_16","alias_value":"ZNSOJVBLENAKSPY5","created_at":"2026-07-05T11:16:12Z"},{"alias_kind":"pith_short_8","alias_value":"ZNSOJVBL","created_at":"2026-07-05T11:16:12Z"}],"graph_snapshots":[{"event_id":"sha256:3db16c7b831b861c8221cfe6cacfb0f2a3b8851e3b5e5d095c7b9c076b000733","target":"graph","created_at":"2026-07-05T11:16:12Z","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/2506.04302/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Red teaming has proven to be an effective method for identifying and mitigating vulnerabilities in Large Language Models (LLMs). Reinforcement Fine-Tuning (RFT) has emerged as a promising strategy among existing red teaming techniques. However, a lack of a unified benchmark hinders current RFT-based red teaming methods. Implementation details, especially in Proximal Policy Optimization (PPO)-based RFT, significantly affect outcome stability and reproducibility. To address this issue, we introduce RedRFT, a lightweight benchmark designed to simplify and standardize the implementation and evalua","authors_text":"Cong Wang, Wei-Bin Lee, Xiang Zheng, Xingjun Ma","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-06-04T14:50:24Z","title":"RedRFT: A Light-Weight Benchmark for Reinforcement Fine-Tuning-Based Red Teaming"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.04302","kind":"arxiv","version":1},"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":"sha256:4584fc0e7454fe7be6379b0d19c7b3c33b5d87dc6481935efebd6ada1dab5683","target":"record","created_at":"2026-07-05T11:16:12Z","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":"2320591b3c37af2b463275423cf074d9257875b055672f0e02e705552fccf002","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-06-04T14:50:24Z","title_canon_sha256":"021dce96743b3204a00e6c691d51f1e4f972935b99a92b8d78bbca861e4297ff"},"schema_version":"1.0","source":{"id":"2506.04302","kind":"arxiv","version":1}},"canonical_sha256":"cb64e4d42b2340a93f1d536350b69167f3419e6812abe397ac50188ad89bf871","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"cb64e4d42b2340a93f1d536350b69167f3419e6812abe397ac50188ad89bf871","first_computed_at":"2026-07-05T11:16:12.051864Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:16:12.051864Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"hqZQiHTGuNoPC9d//JSF5Z7SxBFUmW97GkXuw3YBTzufW8/K1Yee8rN7R/TIZ51773UjVGnucSKYdTzxWhVpAg==","signature_status":"signed_v1","signed_at":"2026-07-05T11:16:12.052371Z","signed_message":"canonical_sha256_bytes"},"source_id":"2506.04302","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4584fc0e7454fe7be6379b0d19c7b3c33b5d87dc6481935efebd6ada1dab5683","sha256:3db16c7b831b861c8221cfe6cacfb0f2a3b8851e3b5e5d095c7b9c076b000733"],"state_sha256":"76ebec0f2416e738a0d83d5741e26006eed71a292037ba91967d67f214694f51"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3Ob6/udF1O+uxf1qTxOs9dy2GY9yQhG0vDaFJNlxYTzIoCDswweK2k9ph9bMENGCijgO+DfeEBrxl8C7FYweAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-19T22:02:37.858767Z","bundle_sha256":"0dbc22de9b4f00f33e96343da3be8bd4f9030796b0cd77e7c0730681a99c92e4"}}