{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:MFSZ5HAVI7Q6CYE63HJ45O57AN","short_pith_number":"pith:MFSZ5HAV","canonical_record":{"source":{"id":"2608.01418","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.AI","submitted_at":"2026-08-02T18:02:03Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"feb6b41014c685fe2a0b1201ce26e1f54e221d69809561aa4057992bd2aa9966","abstract_canon_sha256":"8f39b2caf9ff7db46e830fb06e52581c2aaf01560018e5b25279f4dc9d808612"},"schema_version":"1.0"},"canonical_sha256":"61659e9c1547e1e1609ed9d3cebbbf0356112596c45eee590f9f0302bae80ae7","source":{"kind":"arxiv","id":"2608.01418","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2608.01418","created_at":"2026-08-04T02:05:11Z"},{"alias_kind":"arxiv_version","alias_value":"2608.01418v1","created_at":"2026-08-04T02:05:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.01418","created_at":"2026-08-04T02:05:11Z"},{"alias_kind":"pith_short_12","alias_value":"MFSZ5HAVI7Q6","created_at":"2026-08-04T02:05:11Z"},{"alias_kind":"pith_short_16","alias_value":"MFSZ5HAVI7Q6CYE6","created_at":"2026-08-04T02:05:11Z"},{"alias_kind":"pith_short_8","alias_value":"MFSZ5HAV","created_at":"2026-08-04T02:05:11Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:MFSZ5HAVI7Q6CYE63HJ45O57AN","target":"record","payload":{"canonical_record":{"source":{"id":"2608.01418","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.AI","submitted_at":"2026-08-02T18:02:03Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"feb6b41014c685fe2a0b1201ce26e1f54e221d69809561aa4057992bd2aa9966","abstract_canon_sha256":"8f39b2caf9ff7db46e830fb06e52581c2aaf01560018e5b25279f4dc9d808612"},"schema_version":"1.0"},"canonical_sha256":"61659e9c1547e1e1609ed9d3cebbbf0356112596c45eee590f9f0302bae80ae7","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-04T02:05:11.566362Z","signature_b64":"xw2kbr+gKkHMNbLJqqkB9KLA4ACYNuB6GMgAloYIJiTVOm1ZWtxOY7Lrd1TEW2k2p08OhW7YnaBVZNCGkjQwBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"61659e9c1547e1e1609ed9d3cebbbf0356112596c45eee590f9f0302bae80ae7","last_reissued_at":"2026-08-04T02:05:11.564616Z","signature_status":"signed_v1","first_computed_at":"2026-08-04T02:05:11.564616Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2608.01418","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":"2026-08-04T02:05:11Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"QStz/u9wWIhJfVhKD61/ZKNcPI95V7ETV1owbu6BPqgM8YWw8mokGtj7DRXgt0X1AWRFKCr9bB5bSAR4DL1vAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-11T18:35:26.840179Z"},"content_sha256":"f113e4b36acebb925aa4c7a5954c477b30306207bb2696855d333eea27f8dc60","schema_version":"1.0","event_id":"sha256:f113e4b36acebb925aa4c7a5954c477b30306207bb2696855d333eea27f8dc60"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:MFSZ5HAVI7Q6CYE63HJ45O57AN","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Reusing Rollouts under Policy Lag: Prefix-Normalized Policy Optimization for LLM Reinforcement Learning","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"alantzhao, Haoyang Long, Huan Xiong, jasperawang, Jiahua Yang, Jiali Xu, Lei Jiang, Rui Wang, Wenhao Zhang, Yawei Wang, Yibo Xie, Zibo Yang","submitted_at":"2026-08-02T18:02:03Z","abstract_excerpt":"Autoregressive rollout generation is a major computational cost in reinforcement learning for large language models. Reusing each rollout batch for additional learner updates amortizes this cost, but later updates become increasingly off-policy as the learner departs from the behavior policy. At a token position, exact off-policy correction must account for both the current action and the probability of reaching its prefix. The cumulative importance ratio provides this correction, but its product form can produce an unwieldy dynamic range. We study Prefix-Normalized Policy Optimization (PNPO),"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.01418","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/2608.01418/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":"2026-08-04T02:05:11Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"TPRH4OSyO4FnkJe01nI5gNq72MQjqL4Kla31R30mCRSkABn4Iuk2UEL56/0nQ0yN3EP6FWT814jE/HLCGyy/Bg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-11T18:35:26.840689Z"},"content_sha256":"127ce161231483195222a7d45f270866b85274b59be0d6c7b8c1387f14373ada","schema_version":"1.0","event_id":"sha256:127ce161231483195222a7d45f270866b85274b59be0d6c7b8c1387f14373ada"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/MFSZ5HAVI7Q6CYE63HJ45O57AN/bundle.json","state_url":"https://pith.science/pith/MFSZ5HAVI7Q6CYE63HJ45O57AN/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/MFSZ5HAVI7Q6CYE63HJ45O57AN/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":"2026-08-11T18:35:26Z","links":{"resolver":"https://pith.science/pith/MFSZ5HAVI7Q6CYE63HJ45O57AN","bundle":"https://pith.science/pith/MFSZ5HAVI7Q6CYE63HJ45O57AN/bundle.json","state":"https://pith.science/pith/MFSZ5HAVI7Q6CYE63HJ45O57AN/state.json","well_known_bundle":"https://pith.science/.well-known/pith/MFSZ5HAVI7Q6CYE63HJ45O57AN/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:MFSZ5HAVI7Q6CYE63HJ45O57AN","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":"8f39b2caf9ff7db46e830fb06e52581c2aaf01560018e5b25279f4dc9d808612","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.AI","submitted_at":"2026-08-02T18:02:03Z","title_canon_sha256":"feb6b41014c685fe2a0b1201ce26e1f54e221d69809561aa4057992bd2aa9966"},"schema_version":"1.0","source":{"id":"2608.01418","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2608.01418","created_at":"2026-08-04T02:05:11Z"},{"alias_kind":"arxiv_version","alias_value":"2608.01418v1","created_at":"2026-08-04T02:05:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.01418","created_at":"2026-08-04T02:05:11Z"},{"alias_kind":"pith_short_12","alias_value":"MFSZ5HAVI7Q6","created_at":"2026-08-04T02:05:11Z"},{"alias_kind":"pith_short_16","alias_value":"MFSZ5HAVI7Q6CYE6","created_at":"2026-08-04T02:05:11Z"},{"alias_kind":"pith_short_8","alias_value":"MFSZ5HAV","created_at":"2026-08-04T02:05:11Z"}],"graph_snapshots":[{"event_id":"sha256:127ce161231483195222a7d45f270866b85274b59be0d6c7b8c1387f14373ada","target":"graph","created_at":"2026-08-04T02:05:11Z","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/2608.01418/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Autoregressive rollout generation is a major computational cost in reinforcement learning for large language models. Reusing each rollout batch for additional learner updates amortizes this cost, but later updates become increasingly off-policy as the learner departs from the behavior policy. At a token position, exact off-policy correction must account for both the current action and the probability of reaching its prefix. The cumulative importance ratio provides this correction, but its product form can produce an unwieldy dynamic range. We study Prefix-Normalized Policy Optimization (PNPO),","authors_text":"alantzhao, Haoyang Long, Huan Xiong, jasperawang, Jiahua Yang, Jiali Xu, Lei Jiang, Rui Wang, Wenhao Zhang, Yawei Wang, Yibo Xie, Zibo Yang","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.AI","submitted_at":"2026-08-02T18:02:03Z","title":"Reusing Rollouts under Policy Lag: Prefix-Normalized Policy Optimization for LLM Reinforcement Learning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.01418","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:f113e4b36acebb925aa4c7a5954c477b30306207bb2696855d333eea27f8dc60","target":"record","created_at":"2026-08-04T02:05:11Z","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":"8f39b2caf9ff7db46e830fb06e52581c2aaf01560018e5b25279f4dc9d808612","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.AI","submitted_at":"2026-08-02T18:02:03Z","title_canon_sha256":"feb6b41014c685fe2a0b1201ce26e1f54e221d69809561aa4057992bd2aa9966"},"schema_version":"1.0","source":{"id":"2608.01418","kind":"arxiv","version":1}},"canonical_sha256":"61659e9c1547e1e1609ed9d3cebbbf0356112596c45eee590f9f0302bae80ae7","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"61659e9c1547e1e1609ed9d3cebbbf0356112596c45eee590f9f0302bae80ae7","first_computed_at":"2026-08-04T02:05:11.564616Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-08-04T02:05:11.564616Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"xw2kbr+gKkHMNbLJqqkB9KLA4ACYNuB6GMgAloYIJiTVOm1ZWtxOY7Lrd1TEW2k2p08OhW7YnaBVZNCGkjQwBg==","signature_status":"signed_v1","signed_at":"2026-08-04T02:05:11.566362Z","signed_message":"canonical_sha256_bytes"},"source_id":"2608.01418","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f113e4b36acebb925aa4c7a5954c477b30306207bb2696855d333eea27f8dc60","sha256:127ce161231483195222a7d45f270866b85274b59be0d6c7b8c1387f14373ada"],"state_sha256":"fc625255a99edf7d3354b43f405e4caf5554f8d736ce24a4142a9f1c85de7adf"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1if0FtlLR3TKPBfx3SYPwM/I9yNdQRoWTsC6BxXQntRv97H2T15mLSFENPb79mEKauunHGLAXBj+L9fY8uKvAw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-11T18:35:26.846010Z","bundle_sha256":"a6cc3c7a948182926101bfda90374de25405c3379b621398e07a3a2680c3be5e"}}