{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:6Z6SGOOZBIBMZHMB5M3QWTPNA2","short_pith_number":"pith:6Z6SGOOZ","schema_version":"1.0","canonical_sha256":"f67d2339d90a02cc9d81eb370b4ded069ebb1131243f75b4dbd88b0150171590","source":{"kind":"arxiv","id":"2606.20008","version":1},"attestation_state":"computed","paper":{"title":"VIMPO: Value-Implicit Policy Optimization for LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Aosong Feng, Dawn Song, Sergey Levine, Xuandong Zhao, Zhewei Kang","submitted_at":"2026-06-18T09:44:12Z","abstract_excerpt":"Reinforcement learning with verifiable rewards has become a central tool for improving the reasoning ability of large language models, but current methods face a trade-off between simplicity and credit assignment. Group-relative methods such as GRPO avoid training a critic, but typically assign a trajectory-level advantage to every token. Actor-critic methods provide denser learning signals, but require a learned value function with its own training instability. We introduce VIMPO, a critic-free policy optimization method that derives a policy-implied value function from the optimality conditi"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2606.20008","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-06-18T09:44:12Z","cross_cats_sorted":[],"title_canon_sha256":"b2a44c01ec7db09138c2c61daf2e316cb91e6441afb0e01ac61ab9663faa5bae","abstract_canon_sha256":"76b87c27153b95e3a0e89147aa1f92e16e1a4018b1e8c40f199c55f833538cb2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:13:00.500773Z","signature_b64":"8711B4kobtOD+qNRWDOIcl/t8BC5LBM1LqbS1uqyxirFeIwg0VCEyZP66YHKeY7nm4CqOKrKVaduu/LCrUVWDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f67d2339d90a02cc9d81eb370b4ded069ebb1131243f75b4dbd88b0150171590","last_reissued_at":"2026-06-19T16:13:00.500424Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:13:00.500424Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"VIMPO: Value-Implicit Policy Optimization for LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Aosong Feng, Dawn Song, Sergey Levine, Xuandong Zhao, Zhewei Kang","submitted_at":"2026-06-18T09:44:12Z","abstract_excerpt":"Reinforcement learning with verifiable rewards has become a central tool for improving the reasoning ability of large language models, but current methods face a trade-off between simplicity and credit assignment. Group-relative methods such as GRPO avoid training a critic, but typically assign a trajectory-level advantage to every token. Actor-critic methods provide denser learning signals, but require a learned value function with its own training instability. We introduce VIMPO, a critic-free policy optimization method that derives a policy-implied value function from the optimality conditi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.20008","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/2606.20008/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2606.20008","created_at":"2026-06-19T16:13:00.500485+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.20008v1","created_at":"2026-06-19T16:13:00.500485+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.20008","created_at":"2026-06-19T16:13:00.500485+00:00"},{"alias_kind":"pith_short_12","alias_value":"6Z6SGOOZBIBM","created_at":"2026-06-19T16:13:00.500485+00:00"},{"alias_kind":"pith_short_16","alias_value":"6Z6SGOOZBIBMZHMB","created_at":"2026-06-19T16:13:00.500485+00:00"},{"alias_kind":"pith_short_8","alias_value":"6Z6SGOOZ","created_at":"2026-06-19T16:13:00.500485+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6Z6SGOOZBIBMZHMB5M3QWTPNA2","json":"https://pith.science/pith/6Z6SGOOZBIBMZHMB5M3QWTPNA2.json","graph_json":"https://pith.science/api/pith-number/6Z6SGOOZBIBMZHMB5M3QWTPNA2/graph.json","events_json":"https://pith.science/api/pith-number/6Z6SGOOZBIBMZHMB5M3QWTPNA2/events.json","paper":"https://pith.science/paper/6Z6SGOOZ"},"agent_actions":{"view_html":"https://pith.science/pith/6Z6SGOOZBIBMZHMB5M3QWTPNA2","download_json":"https://pith.science/pith/6Z6SGOOZBIBMZHMB5M3QWTPNA2.json","view_paper":"https://pith.science/paper/6Z6SGOOZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.20008&json=true","fetch_graph":"https://pith.science/api/pith-number/6Z6SGOOZBIBMZHMB5M3QWTPNA2/graph.json","fetch_events":"https://pith.science/api/pith-number/6Z6SGOOZBIBMZHMB5M3QWTPNA2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6Z6SGOOZBIBMZHMB5M3QWTPNA2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6Z6SGOOZBIBMZHMB5M3QWTPNA2/action/storage_attestation","attest_author":"https://pith.science/pith/6Z6SGOOZBIBMZHMB5M3QWTPNA2/action/author_attestation","sign_citation":"https://pith.science/pith/6Z6SGOOZBIBMZHMB5M3QWTPNA2/action/citation_signature","submit_replication":"https://pith.science/pith/6Z6SGOOZBIBMZHMB5M3QWTPNA2/action/replication_record"}},"created_at":"2026-06-19T16:13:00.500485+00:00","updated_at":"2026-06-19T16:13:00.500485+00:00"}