{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:V7QZ7YMIFHEJJ6KNNL7K5JGARV","short_pith_number":"pith:V7QZ7YMI","canonical_record":{"source":{"id":"2508.09074","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-12T16:49:18Z","cross_cats_sorted":[],"title_canon_sha256":"4718fadb8d8e2bdc05066fe5eb3edc9127b2833fadb9c328aba4d272dc64adc8","abstract_canon_sha256":"5c53ed221ccc547184effa5d80237b09960f693423f36c4027b6e9701f487987"},"schema_version":"1.0"},"canonical_sha256":"afe19fe18829c894f94d6afeaea4c08d623d506532645404e533950e4f41d94b","source":{"kind":"arxiv","id":"2508.09074","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.09074","created_at":"2026-07-05T11:52:39Z"},{"alias_kind":"arxiv_version","alias_value":"2508.09074v1","created_at":"2026-07-05T11:52:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.09074","created_at":"2026-07-05T11:52:39Z"},{"alias_kind":"pith_short_12","alias_value":"V7QZ7YMIFHEJ","created_at":"2026-07-05T11:52:39Z"},{"alias_kind":"pith_short_16","alias_value":"V7QZ7YMIFHEJJ6KN","created_at":"2026-07-05T11:52:39Z"},{"alias_kind":"pith_short_8","alias_value":"V7QZ7YMI","created_at":"2026-07-05T11:52:39Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:V7QZ7YMIFHEJJ6KNNL7K5JGARV","target":"record","payload":{"canonical_record":{"source":{"id":"2508.09074","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-12T16:49:18Z","cross_cats_sorted":[],"title_canon_sha256":"4718fadb8d8e2bdc05066fe5eb3edc9127b2833fadb9c328aba4d272dc64adc8","abstract_canon_sha256":"5c53ed221ccc547184effa5d80237b09960f693423f36c4027b6e9701f487987"},"schema_version":"1.0"},"canonical_sha256":"afe19fe18829c894f94d6afeaea4c08d623d506532645404e533950e4f41d94b","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:52:39.412326Z","signature_b64":"pZaccQO7LhnF5s5aM+iSYLxEzvIrdPNc8QSDm6PVtzO647W0UT3ZUx2QXvNcTDvc44w4DdWL8SxE2bjcHVj3Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"afe19fe18829c894f94d6afeaea4c08d623d506532645404e533950e4f41d94b","last_reissued_at":"2026-07-05T11:52:39.411755Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:52:39.411755Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2508.09074","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":"2026-07-05T11:52:39Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"th7260N/+gceRFJacV9rTvqxysYp2NQ7R1l3C2/GnG3VnU3RMZ25TzSDrGNf8ydM35nUeDyW7jRjwO9UciZBAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-20T11:54:18.730794Z"},"content_sha256":"2db11db1e4ab31ed25762565874734b86ec92f20515a99e5d5a42c667842ae8e","schema_version":"1.0","event_id":"sha256:2db11db1e4ab31ed25762565874734b86ec92f20515a99e5d5a42c667842ae8e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:V7QZ7YMIFHEJJ6KNNL7K5JGARV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"CPO: Addressing Reward Ambiguity in Role-playing Dialogue via Comparative Policy Optimization","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Fei Huang, Feiteng Fang, Rui Wang, Victor Ma, Xinge Ye, Yongbin Li, Yuchuan Wu","submitted_at":"2025-08-12T16:49:18Z","abstract_excerpt":"Reinforcement Learning Fine-Tuning (RLFT) has achieved notable success in tasks with objectively verifiable answers (e.g., code generation, mathematical reasoning), yet struggles with open-ended subjective tasks like role-playing dialogue. Traditional reward modeling approaches, which rely on independent sample-wise scoring, face dual challenges: subjective evaluation criteria and unstable reward signals.Motivated by the insight that human evaluation inherently combines explicit criteria with implicit comparative judgments, we propose Comparative Policy Optimization (CPO). CPO redefines the re"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.09074","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/2508.09074/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":"2026-07-05T11:52:39Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"0pKSWT5qK8hBIO6sxWW6c8P/fAY7dqbhRvYqCzgPCduAsPSC/6MJRqoT8CQN97R13FY6566T/wjFYdD1wbxNBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-20T11:54:18.732062Z"},"content_sha256":"df1008f058f0f2487062587e4b77a5c2a1f2c068b214feb10066c5f5faef11bf","schema_version":"1.0","event_id":"sha256:df1008f058f0f2487062587e4b77a5c2a1f2c068b214feb10066c5f5faef11bf"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/V7QZ7YMIFHEJJ6KNNL7K5JGARV/bundle.json","state_url":"https://pith.science/pith/V7QZ7YMIFHEJJ6KNNL7K5JGARV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/V7QZ7YMIFHEJJ6KNNL7K5JGARV/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":"2026-08-20T11:54:18Z","links":{"resolver":"https://pith.science/pith/V7QZ7YMIFHEJJ6KNNL7K5JGARV","bundle":"https://pith.science/pith/V7QZ7YMIFHEJJ6KNNL7K5JGARV/bundle.json","state":"https://pith.science/pith/V7QZ7YMIFHEJJ6KNNL7K5JGARV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/V7QZ7YMIFHEJJ6KNNL7K5JGARV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:V7QZ7YMIFHEJJ6KNNL7K5JGARV","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":"5c53ed221ccc547184effa5d80237b09960f693423f36c4027b6e9701f487987","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-12T16:49:18Z","title_canon_sha256":"4718fadb8d8e2bdc05066fe5eb3edc9127b2833fadb9c328aba4d272dc64adc8"},"schema_version":"1.0","source":{"id":"2508.09074","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.09074","created_at":"2026-07-05T11:52:39Z"},{"alias_kind":"arxiv_version","alias_value":"2508.09074v1","created_at":"2026-07-05T11:52:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.09074","created_at":"2026-07-05T11:52:39Z"},{"alias_kind":"pith_short_12","alias_value":"V7QZ7YMIFHEJ","created_at":"2026-07-05T11:52:39Z"},{"alias_kind":"pith_short_16","alias_value":"V7QZ7YMIFHEJJ6KN","created_at":"2026-07-05T11:52:39Z"},{"alias_kind":"pith_short_8","alias_value":"V7QZ7YMI","created_at":"2026-07-05T11:52:39Z"}],"graph_snapshots":[{"event_id":"sha256:df1008f058f0f2487062587e4b77a5c2a1f2c068b214feb10066c5f5faef11bf","target":"graph","created_at":"2026-07-05T11:52:39Z","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/2508.09074/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Reinforcement Learning Fine-Tuning (RLFT) has achieved notable success in tasks with objectively verifiable answers (e.g., code generation, mathematical reasoning), yet struggles with open-ended subjective tasks like role-playing dialogue. Traditional reward modeling approaches, which rely on independent sample-wise scoring, face dual challenges: subjective evaluation criteria and unstable reward signals.Motivated by the insight that human evaluation inherently combines explicit criteria with implicit comparative judgments, we propose Comparative Policy Optimization (CPO). CPO redefines the re","authors_text":"Fei Huang, Feiteng Fang, Rui Wang, Victor Ma, Xinge Ye, Yongbin Li, Yuchuan Wu","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-12T16:49:18Z","title":"CPO: Addressing Reward Ambiguity in Role-playing Dialogue via Comparative Policy Optimization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.09074","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":"sha256:2db11db1e4ab31ed25762565874734b86ec92f20515a99e5d5a42c667842ae8e","target":"record","created_at":"2026-07-05T11:52:39Z","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":"5c53ed221ccc547184effa5d80237b09960f693423f36c4027b6e9701f487987","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-12T16:49:18Z","title_canon_sha256":"4718fadb8d8e2bdc05066fe5eb3edc9127b2833fadb9c328aba4d272dc64adc8"},"schema_version":"1.0","source":{"id":"2508.09074","kind":"arxiv","version":1}},"canonical_sha256":"afe19fe18829c894f94d6afeaea4c08d623d506532645404e533950e4f41d94b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"afe19fe18829c894f94d6afeaea4c08d623d506532645404e533950e4f41d94b","first_computed_at":"2026-07-05T11:52:39.411755Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:52:39.411755Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"pZaccQO7LhnF5s5aM+iSYLxEzvIrdPNc8QSDm6PVtzO647W0UT3ZUx2QXvNcTDvc44w4DdWL8SxE2bjcHVj3Ag==","signature_status":"signed_v1","signed_at":"2026-07-05T11:52:39.412326Z","signed_message":"canonical_sha256_bytes"},"source_id":"2508.09074","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:2db11db1e4ab31ed25762565874734b86ec92f20515a99e5d5a42c667842ae8e","sha256:df1008f058f0f2487062587e4b77a5c2a1f2c068b214feb10066c5f5faef11bf"],"state_sha256":"b0d16a7ff3647fd0e8696bc15a03e47bbfee2d3d328bfc4c58fcd8da9be52d44"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"T1gSWQ3cA3is6GBncgsYw41+kEcByp5GRcni8NeAJZWH85i7VyRDVD6QLPOZoAg5TTWxljJgHMkVnGVaSuWmAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-20T11:54:18.739843Z","bundle_sha256":"e5114dbf4030f69efc97e7fe94f51f87622b9645d941bd42b8f4ae6b1c4ead2c"}}