{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:JCXXAV75TATT42ZK5WXRYP3THM","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":"0df58d1f7ae56d477af525c215601a54165e25616c4810facc5fbdb690ce81bf","cross_cats_sorted":["cs.AI","cs.CL","stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-04-29T17:58:30Z","title_canon_sha256":"85ccbc00c61d4745b1680d32b211e5c76eedec156dd5586127798358768fff8b"},"schema_version":"1.0","source":{"id":"2404.18922","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.18922","created_at":"2026-07-05T11:06:20Z"},{"alias_kind":"arxiv_version","alias_value":"2404.18922v4","created_at":"2026-07-05T11:06:20Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.18922","created_at":"2026-07-05T11:06:20Z"},{"alias_kind":"pith_short_12","alias_value":"JCXXAV75TATT","created_at":"2026-07-05T11:06:20Z"},{"alias_kind":"pith_short_16","alias_value":"JCXXAV75TATT42ZK","created_at":"2026-07-05T11:06:20Z"},{"alias_kind":"pith_short_8","alias_value":"JCXXAV75","created_at":"2026-07-05T11:06:20Z"}],"graph_snapshots":[{"event_id":"sha256:b8d55a8912fd6de0475b4687734e50c9c7cb5eef5b0b757592fd489dfea944c1","target":"graph","created_at":"2026-07-05T11:06:20Z","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/2404.18922/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"In the classical Reinforcement Learning from Human Feedback (RLHF) framework, Proximal Policy Optimization (PPO) is employed to learn from sparse, sentence-level rewards -- a challenging scenario in traditional deep reinforcement learning. Despite the great successes of PPO in the alignment of large language models, its open-source implementation is still largely sub-optimal. To address these issues, we introduce a framework that models RLHF problems as a Markov decision process (MDP), enabling the capture of fine-grained token-wise information. Under this framework, we introduce an algorithm ","authors_text":"Di He, Guhao Feng, Han Zhong, Jiang Bian, Liwei Wang, Li Zhao, Wei Xiong, Xinle Cheng, Zikang Shan","cross_cats":["cs.AI","cs.CL","stat.ML"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-04-29T17:58:30Z","title":"DPO Meets PPO: Reinforced Token Optimization for RLHF"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.18922","kind":"arxiv","version":4},"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:f5a50b9a6b80a9eeae85b102f301c6b7821d9880312ad93276e880f7ffa082e6","target":"record","created_at":"2026-07-05T11:06:20Z","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":"0df58d1f7ae56d477af525c215601a54165e25616c4810facc5fbdb690ce81bf","cross_cats_sorted":["cs.AI","cs.CL","stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-04-29T17:58:30Z","title_canon_sha256":"85ccbc00c61d4745b1680d32b211e5c76eedec156dd5586127798358768fff8b"},"schema_version":"1.0","source":{"id":"2404.18922","kind":"arxiv","version":4}},"canonical_sha256":"48af7057fd98273e6b2aedaf1c3f733b3d60f86ba44f1bb3997f99f604eac380","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"48af7057fd98273e6b2aedaf1c3f733b3d60f86ba44f1bb3997f99f604eac380","first_computed_at":"2026-07-05T11:06:20.576721Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:06:20.576721Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"TjxHo3zh8w4IN6MG7RQ121MMP/DGPEFxINWsqvcm4XViOeBvQP6W1yEoTaMXeDAx1XDNLB29GrpLyZEp39+yDg==","signature_status":"signed_v1","signed_at":"2026-07-05T11:06:20.577252Z","signed_message":"canonical_sha256_bytes"},"source_id":"2404.18922","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f5a50b9a6b80a9eeae85b102f301c6b7821d9880312ad93276e880f7ffa082e6","sha256:b8d55a8912fd6de0475b4687734e50c9c7cb5eef5b0b757592fd489dfea944c1"],"state_sha256":"3149e8aec2fd44eda3595b3b3ab5aa508339871102a1d4d165347d478484ab34"}