{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:JPCERWN6CSYP52FQI6JVAO5MER","short_pith_number":"pith:JPCERWN6","canonical_record":{"source":{"id":"2501.15453","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-26T08:49:46Z","cross_cats_sorted":[],"title_canon_sha256":"74f61fd090e57908da7559420c3dffcf55c0d931ab6cbabe9f557ad890b57bb5","abstract_canon_sha256":"a49f6afcc2c491ae30d85669963172520d022e0a25064456c9b261dd1c379f96"},"schema_version":"1.0"},"canonical_sha256":"4bc448d9be14b0fee8b04793503bac24729b3de8a120791c79b19346afc4bbaa","source":{"kind":"arxiv","id":"2501.15453","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.15453","created_at":"2026-07-05T10:06:11Z"},{"alias_kind":"arxiv_version","alias_value":"2501.15453v2","created_at":"2026-07-05T10:06:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.15453","created_at":"2026-07-05T10:06:11Z"},{"alias_kind":"pith_short_12","alias_value":"JPCERWN6CSYP","created_at":"2026-07-05T10:06:11Z"},{"alias_kind":"pith_short_16","alias_value":"JPCERWN6CSYP52FQ","created_at":"2026-07-05T10:06:11Z"},{"alias_kind":"pith_short_8","alias_value":"JPCERWN6","created_at":"2026-07-05T10:06:11Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:JPCERWN6CSYP52FQI6JVAO5MER","target":"record","payload":{"canonical_record":{"source":{"id":"2501.15453","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-26T08:49:46Z","cross_cats_sorted":[],"title_canon_sha256":"74f61fd090e57908da7559420c3dffcf55c0d931ab6cbabe9f557ad890b57bb5","abstract_canon_sha256":"a49f6afcc2c491ae30d85669963172520d022e0a25064456c9b261dd1c379f96"},"schema_version":"1.0"},"canonical_sha256":"4bc448d9be14b0fee8b04793503bac24729b3de8a120791c79b19346afc4bbaa","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:06:11.049524Z","signature_b64":"Ur3Pln0Nu/Ea8EQizahuaRUdh8DaPD4l3YRVaVT+eaWknOWpe8Yyp86DF4oHEWvrGfs7aFK0pAGeHtYmDxhJAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4bc448d9be14b0fee8b04793503bac24729b3de8a120791c79b19346afc4bbaa","last_reissued_at":"2026-07-05T10:06:11.048965Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:06:11.048965Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2501.15453","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":"2026-07-05T10:06:11Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"O8HW4oMW5IxKUr/EbSuaNFuuyfSyKUKbC/U2xFO92iYqXBAeHo6OYjxUNBIy7QDM34Qb1C/pO3SfII1C3vZfAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T21:31:05.945309Z"},"content_sha256":"0e91736b84df1dbf81a2bf44f57a05390c87034629b8b4be04b6afd27688104f","schema_version":"1.0","event_id":"sha256:0e91736b84df1dbf81a2bf44f57a05390c87034629b8b4be04b6afd27688104f"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:JPCERWN6CSYP52FQI6JVAO5MER","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Data-adaptive Safety Rules for Training Reward Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jingxuan Fan, Mingye Gao, Weiyu Li, Xiaomin Li, Zhiwei Zhang","submitted_at":"2025-01-26T08:49:46Z","abstract_excerpt":"Reinforcement Learning from Human Feedback (RLHF) is commonly employed to tailor models to human preferences, especially to improve the safety of outputs from large language models (LLMs). Traditionally, this method depends on selecting preferred responses from pairs. However, due to the variability in human opinions and the challenges in directly comparing two responses, there is an increasing trend towards fine-grained annotation approaches that evaluate responses using multiple targeted metrics or rules. The challenge lies in efficiently choosing and applying these rules to handle the diver"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.15453","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/2501.15453/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":"2026-07-05T10:06:11Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"GpYeqIZlpesE7W5e8N2Fb7LUimqj5XHm/ZiD49StlnZ6Jn+gaMCgPQXwHzTcreZp/LzVgNQtIYP+0Ru13WzPCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T21:31:05.945798Z"},"content_sha256":"2a10377bc15af8966698e189a35cf514a2c5871bb649d2a0b72e3c3dd4919162","schema_version":"1.0","event_id":"sha256:2a10377bc15af8966698e189a35cf514a2c5871bb649d2a0b72e3c3dd4919162"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/JPCERWN6CSYP52FQI6JVAO5MER/bundle.json","state_url":"https://pith.science/pith/JPCERWN6CSYP52FQI6JVAO5MER/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/JPCERWN6CSYP52FQI6JVAO5MER/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":"2026-08-06T21:31:05Z","links":{"resolver":"https://pith.science/pith/JPCERWN6CSYP52FQI6JVAO5MER","bundle":"https://pith.science/pith/JPCERWN6CSYP52FQI6JVAO5MER/bundle.json","state":"https://pith.science/pith/JPCERWN6CSYP52FQI6JVAO5MER/state.json","well_known_bundle":"https://pith.science/.well-known/pith/JPCERWN6CSYP52FQI6JVAO5MER/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:JPCERWN6CSYP52FQI6JVAO5MER","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":"a49f6afcc2c491ae30d85669963172520d022e0a25064456c9b261dd1c379f96","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-26T08:49:46Z","title_canon_sha256":"74f61fd090e57908da7559420c3dffcf55c0d931ab6cbabe9f557ad890b57bb5"},"schema_version":"1.0","source":{"id":"2501.15453","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.15453","created_at":"2026-07-05T10:06:11Z"},{"alias_kind":"arxiv_version","alias_value":"2501.15453v2","created_at":"2026-07-05T10:06:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.15453","created_at":"2026-07-05T10:06:11Z"},{"alias_kind":"pith_short_12","alias_value":"JPCERWN6CSYP","created_at":"2026-07-05T10:06:11Z"},{"alias_kind":"pith_short_16","alias_value":"JPCERWN6CSYP52FQ","created_at":"2026-07-05T10:06:11Z"},{"alias_kind":"pith_short_8","alias_value":"JPCERWN6","created_at":"2026-07-05T10:06:11Z"}],"graph_snapshots":[{"event_id":"sha256:2a10377bc15af8966698e189a35cf514a2c5871bb649d2a0b72e3c3dd4919162","target":"graph","created_at":"2026-07-05T10:06:11Z","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/2501.15453/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Reinforcement Learning from Human Feedback (RLHF) is commonly employed to tailor models to human preferences, especially to improve the safety of outputs from large language models (LLMs). Traditionally, this method depends on selecting preferred responses from pairs. However, due to the variability in human opinions and the challenges in directly comparing two responses, there is an increasing trend towards fine-grained annotation approaches that evaluate responses using multiple targeted metrics or rules. The challenge lies in efficiently choosing and applying these rules to handle the diver","authors_text":"Jingxuan Fan, Mingye Gao, Weiyu Li, Xiaomin Li, Zhiwei Zhang","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-26T08:49:46Z","title":"Data-adaptive Safety Rules for Training Reward Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.15453","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":"sha256:0e91736b84df1dbf81a2bf44f57a05390c87034629b8b4be04b6afd27688104f","target":"record","created_at":"2026-07-05T10:06:11Z","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":"a49f6afcc2c491ae30d85669963172520d022e0a25064456c9b261dd1c379f96","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-26T08:49:46Z","title_canon_sha256":"74f61fd090e57908da7559420c3dffcf55c0d931ab6cbabe9f557ad890b57bb5"},"schema_version":"1.0","source":{"id":"2501.15453","kind":"arxiv","version":2}},"canonical_sha256":"4bc448d9be14b0fee8b04793503bac24729b3de8a120791c79b19346afc4bbaa","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4bc448d9be14b0fee8b04793503bac24729b3de8a120791c79b19346afc4bbaa","first_computed_at":"2026-07-05T10:06:11.048965Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:06:11.048965Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Ur3Pln0Nu/Ea8EQizahuaRUdh8DaPD4l3YRVaVT+eaWknOWpe8Yyp86DF4oHEWvrGfs7aFK0pAGeHtYmDxhJAg==","signature_status":"signed_v1","signed_at":"2026-07-05T10:06:11.049524Z","signed_message":"canonical_sha256_bytes"},"source_id":"2501.15453","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0e91736b84df1dbf81a2bf44f57a05390c87034629b8b4be04b6afd27688104f","sha256:2a10377bc15af8966698e189a35cf514a2c5871bb649d2a0b72e3c3dd4919162"],"state_sha256":"bd46c5019112df953af4328100c81ebeb3016f168dd2194bfa0e553b7877a8c3"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"VBvatnqU/4wbyt1bIEQjmFdqmDBmXEp1l75pt9DY1koT9CMKB4Qnz0ZqPRQWD93E+sISy4SBJOXiKW8he/eMCw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-06T21:31:05.949366Z","bundle_sha256":"826b4d8622d7e842af4a9d5da708e22212a39697afb053a750af1c18c1d72e3f"}}