{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:PBHJ67NDZ5N3WUQMWZZZNKTZQA","short_pith_number":"pith:PBHJ67ND","canonical_record":{"source":{"id":"2602.08503","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-02-09T10:55:13Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"1df94f5bb4338f09c6d55d2ff6160db76f35e3ba60c6054c11470b2fc0625955","abstract_canon_sha256":"efd48096fd33e97e88e3f8fefa35fc389d623f0fafe0acd2980a8c7ec510eed1"},"schema_version":"1.0"},"canonical_sha256":"784e9f7da3cf5bbb520cb67396aa79803eeedd6dfd78dc53f0fe614384ff2d39","source":{"kind":"arxiv","id":"2602.08503","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.08503","created_at":"2026-06-05T01:14:34Z"},{"alias_kind":"arxiv_version","alias_value":"2602.08503v2","created_at":"2026-06-05T01:14:34Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.08503","created_at":"2026-06-05T01:14:34Z"},{"alias_kind":"pith_short_12","alias_value":"PBHJ67NDZ5N3","created_at":"2026-06-05T01:14:34Z"},{"alias_kind":"pith_short_16","alias_value":"PBHJ67NDZ5N3WUQM","created_at":"2026-06-05T01:14:34Z"},{"alias_kind":"pith_short_8","alias_value":"PBHJ67ND","created_at":"2026-06-05T01:14:34Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:PBHJ67NDZ5N3WUQMWZZZNKTZQA","target":"record","payload":{"canonical_record":{"source":{"id":"2602.08503","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-02-09T10:55:13Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"1df94f5bb4338f09c6d55d2ff6160db76f35e3ba60c6054c11470b2fc0625955","abstract_canon_sha256":"efd48096fd33e97e88e3f8fefa35fc389d623f0fafe0acd2980a8c7ec510eed1"},"schema_version":"1.0"},"canonical_sha256":"784e9f7da3cf5bbb520cb67396aa79803eeedd6dfd78dc53f0fe614384ff2d39","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-05T01:14:34.866766Z","signature_b64":"oe+mwC+gS10TU3MvUuAFM0TuYSNQW1KRbVrcXGYfMbhakHuvsVD2W3G03iwtSJhdoOHrYKJGg+FeOWD1VpFPCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"784e9f7da3cf5bbb520cb67396aa79803eeedd6dfd78dc53f0fe614384ff2d39","last_reissued_at":"2026-06-05T01:14:34.865983Z","signature_status":"signed_v1","first_computed_at":"2026-06-05T01:14:34.865983Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2602.08503","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":"2026-06-05T01:14:34Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qMSZpBOWiU9mqnBuOMn+I4Rio8AA37+44TuMzWkRjSg3RnQExY7XeM5vsC4WLCPagcpEHLY1FKk+m/aVkCsUCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-21T04:27:30.158798Z"},"content_sha256":"55de6aa09f9a518e402bbfaff27c820b4b7f1346a2cba9955060d6c9f9089e13","schema_version":"1.0","event_id":"sha256:55de6aa09f9a518e402bbfaff27c820b4b7f1346a2cba9955060d6c9f9089e13"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:PBHJ67NDZ5N3WUQMWZZZNKTZQA","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Learning Self-Correction in Vision-Language Models via Rollout Augmentation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Bolian Li, Ruqi Zhang, Yi Ding, Ziliang Qiu","submitted_at":"2026-02-09T10:55:13Z","abstract_excerpt":"Self-correction is essential for solving complex reasoning problems in vision-language models (VLMs). However, existing reinforcement learning (RL) methods struggle to learn it, as effective self-correction behaviors emerge only rarely, making learning signals extremely sparse. To address this challenge, we propose correction-specific rollouts (Octopus), an RL rollout augmentation framework that synthesizes dense self-correction examples by recombining existing rollouts. This augmentation simultaneously improves sample efficiency due to rollout reuse and stabilizes RL optimization through bala"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.08503","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/2602.08503/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":"2026-06-05T01:14:34Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ruSAaCMJEj2/4CUNTtMjkoEXvO5I/OlGvETOGSlzKSMjA5PlYKXKjBtSkoSwjE+VMIicCMuvfw0d4JuZC5yuDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-21T04:27:30.159295Z"},"content_sha256":"4b27df67e75e71e89d4fe58fab151a3f95b9091d872fc7c49baeba7e24d777bb","schema_version":"1.0","event_id":"sha256:4b27df67e75e71e89d4fe58fab151a3f95b9091d872fc7c49baeba7e24d777bb"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/PBHJ67NDZ5N3WUQMWZZZNKTZQA/bundle.json","state_url":"https://pith.science/pith/PBHJ67NDZ5N3WUQMWZZZNKTZQA/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/PBHJ67NDZ5N3WUQMWZZZNKTZQA/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":"2026-08-21T04:27:30Z","links":{"resolver":"https://pith.science/pith/PBHJ67NDZ5N3WUQMWZZZNKTZQA","bundle":"https://pith.science/pith/PBHJ67NDZ5N3WUQMWZZZNKTZQA/bundle.json","state":"https://pith.science/pith/PBHJ67NDZ5N3WUQMWZZZNKTZQA/state.json","well_known_bundle":"https://pith.science/.well-known/pith/PBHJ67NDZ5N3WUQMWZZZNKTZQA/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:PBHJ67NDZ5N3WUQMWZZZNKTZQA","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":"efd48096fd33e97e88e3f8fefa35fc389d623f0fafe0acd2980a8c7ec510eed1","cross_cats_sorted":["cs.CL","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-02-09T10:55:13Z","title_canon_sha256":"1df94f5bb4338f09c6d55d2ff6160db76f35e3ba60c6054c11470b2fc0625955"},"schema_version":"1.0","source":{"id":"2602.08503","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.08503","created_at":"2026-06-05T01:14:34Z"},{"alias_kind":"arxiv_version","alias_value":"2602.08503v2","created_at":"2026-06-05T01:14:34Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.08503","created_at":"2026-06-05T01:14:34Z"},{"alias_kind":"pith_short_12","alias_value":"PBHJ67NDZ5N3","created_at":"2026-06-05T01:14:34Z"},{"alias_kind":"pith_short_16","alias_value":"PBHJ67NDZ5N3WUQM","created_at":"2026-06-05T01:14:34Z"},{"alias_kind":"pith_short_8","alias_value":"PBHJ67ND","created_at":"2026-06-05T01:14:34Z"}],"graph_snapshots":[{"event_id":"sha256:4b27df67e75e71e89d4fe58fab151a3f95b9091d872fc7c49baeba7e24d777bb","target":"graph","created_at":"2026-06-05T01:14:34Z","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/2602.08503/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Self-correction is essential for solving complex reasoning problems in vision-language models (VLMs). However, existing reinforcement learning (RL) methods struggle to learn it, as effective self-correction behaviors emerge only rarely, making learning signals extremely sparse. To address this challenge, we propose correction-specific rollouts (Octopus), an RL rollout augmentation framework that synthesizes dense self-correction examples by recombining existing rollouts. This augmentation simultaneously improves sample efficiency due to rollout reuse and stabilizes RL optimization through bala","authors_text":"Bolian Li, Ruqi Zhang, Yi Ding, Ziliang Qiu","cross_cats":["cs.CL","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-02-09T10:55:13Z","title":"Learning Self-Correction in Vision-Language Models via Rollout Augmentation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.08503","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":"sha256:55de6aa09f9a518e402bbfaff27c820b4b7f1346a2cba9955060d6c9f9089e13","target":"record","created_at":"2026-06-05T01:14:34Z","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":"efd48096fd33e97e88e3f8fefa35fc389d623f0fafe0acd2980a8c7ec510eed1","cross_cats_sorted":["cs.CL","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-02-09T10:55:13Z","title_canon_sha256":"1df94f5bb4338f09c6d55d2ff6160db76f35e3ba60c6054c11470b2fc0625955"},"schema_version":"1.0","source":{"id":"2602.08503","kind":"arxiv","version":2}},"canonical_sha256":"784e9f7da3cf5bbb520cb67396aa79803eeedd6dfd78dc53f0fe614384ff2d39","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"784e9f7da3cf5bbb520cb67396aa79803eeedd6dfd78dc53f0fe614384ff2d39","first_computed_at":"2026-06-05T01:14:34.865983Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-05T01:14:34.865983Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"oe+mwC+gS10TU3MvUuAFM0TuYSNQW1KRbVrcXGYfMbhakHuvsVD2W3G03iwtSJhdoOHrYKJGg+FeOWD1VpFPCg==","signature_status":"signed_v1","signed_at":"2026-06-05T01:14:34.866766Z","signed_message":"canonical_sha256_bytes"},"source_id":"2602.08503","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:55de6aa09f9a518e402bbfaff27c820b4b7f1346a2cba9955060d6c9f9089e13","sha256:4b27df67e75e71e89d4fe58fab151a3f95b9091d872fc7c49baeba7e24d777bb"],"state_sha256":"7a7f300481df137fe8f08f1f8c5a08d933ed903ae202242b30ba9f743e580fc6"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"jMe+E/6QU+6ANySOXvyVS4fTAvHGWmZZ/l5sQ6RiPUXm7FQBDlerWC41d0CV5Wlld7vRZVIzVLHTgX5F3nj2Dg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-21T04:27:30.164488Z","bundle_sha256":"79599deece101354ef0795dbcc2b2c946bf679f817722029b2e0054cf5997483"}}