{"_canonicalization":{"envelope_id":"axm_ + sha256(envelope minus {signature, axiom_id, anchors})","envelope_signature":"ed25519(envelope minus {signature, axiom_id})","json":"sort_keys=True, separators=(',',':'), ensure_ascii=False, allow_nan=False, utf-8","leaf_hash":"sha256(0x00 || canonical_json(envelope_full))","seal_signature":"ed25519(seal minus {signature, sig_algorithm})"},"axiom_id":"axm_b1149fd71285f1c8af42d871df937e1ec6b6578cf0629be3d43b5298b8a49fc7","bitcoin_anchor":{"bitcoin_attestations":[],"calendar_attestations":[],"ots_url":"","stamped_at":"","status":"pending_next_stamp"},"envelope":{"anchors":[{"chain":"crovia.axiom_graph","height":0,"merkle_proof":"spider_vendor_press_v1","root_at_anchor":"spider_vendor_press_v1"}],"axiom_id":"axm_b1149fd71285f1c8af42d871df937e1ec6b6578cf0629be3d43b5298b8a49fc7","axiom_type":"AX.OBS","body":{"axiom_subtype":"news.vendor_press.v1","category":"news","fingerprint":"92a9b678b001303bc35403e468ead9d442e0e3844671b1826a337a82398b6438","published":"Fri, 05 Jun 2026 00:00:00 -0400","receipt_hash":"92a9b678b001303bc35403e468ead9d442e0e3844671b1826a337a82398b6438","schema":"spider.news.vendor_press.v1","spider":"vendor_press","spider_record":{"axiom_subtype":"news.vendor_press.v1","category":"news","decision_hint":"POSITIVE","envelope_target":"AX.OBS","fingerprint":"92a9b678b001303bc35403e468ead9d442e0e3844671b1826a337a82398b6438","observed_at":"2026-06-05T04:43:18.903985Z","parent_run_hash":"a095e4b5fb6e15f54794b30e69003f3092b5cccd96d935cb910933b96e289144","published":"Fri, 05 Jun 2026 00:00:00 -0400","runtime_version":"0.1.0","schema":"spider.news.vendor_press.v1","source_status":200,"source_url":"https://export.arxiv.org/rss/cs.AI","spider":"vendor_press","summary_excerpt":"arXiv:2606.04560v2 Announce Type: cross \nAbstract: Reinforcement learning from verifiable rewards with GRPO is a standard approach for post-training reasoning LLMs. It remains sample inefficient. Each rollout is used for a single gradient update and then discarded. Naive replay is not well suited in this setting because LLM policies drift quickly per gradient step. Stored rollouts therefore become stale and can destabilize training. We propose a rollout-level replay buffer for GRPO that stores and samples individual rollouts rather than whole groups. The buffer bounds staleness through age eviction. Any rollout older than tau_max training steps is removed. The buffer also preserves on-policy data via fresh-anchored composition. Each batch keeps its fresh on-policy rollouts and then concatenates replay rollouts drawn separately from the buffer. We prioritize replay by per-rollout advantage magnitude and recycle individual rollouts whose advantages are large. Across three Qwen3-Base scal","title":"Rollout-Level Advantage-Prioritized Experience Replay for GRPO","url":"https://arxiv.org/abs/2606.04560","vendor":"arxiv_cs_ai"},"summary":"arXiv:2606.04560v2 Announce Type: cross \nAbstract: Reinforcement learning from verifiable rewards with GRPO is a standard approach for post-training reasoning LLMs. It remains sample inefficient. Each rollout is used for a single gradient update and then discarded. Naive replay is not well suited in this setting because LLM policies drift quickly per gradient step. Stored rollouts therefore become stale and can destabilize training. We propose a rollout-level replay buffer for GRPO that stores and samples individual rollouts rather than whole groups. The buffer bounds staleness through age eviction. Any rollout older than tau_max training steps is removed. The buffer also preserves on-policy data via fresh-anchored composition. Each batch keeps its fresh on-policy rollouts and then concatenates replay rollouts drawn separately from the buffer. We prioritize replay by per-rollout advantage magnitude and recycle individual rollouts whose advantages are large. Across three Qwen3-Base scal","title":"Rollout-Level Advantage-Prioritized Experience Replay for GRPO","vendor":"arxiv_cs_ai"},"confidence":{"method":"deterministic"},"decision":"POSITIVE","issued_at":"2026-06-05T04:43:18Z","notes":"Spider vendor_press (news) news.vendor_press.v1","object":{"captured_by":"crovia.spider.vendor_press","primary_source_url":"https://arxiv.org/abs/2606.04560"},"predecessors":[],"schema":"crovia.axiom.v1","signature":"ed25519:0401051c933b55348bf8e5cee9640d86991f22f0e96b8e2e5805a9d6ab769e2a461eb21b74d2de23e00396607f4ec76b8ab0702b65bf1c639e0e4e3a4bda3e0c","signer":"crovia.substrate","subject":{"observed_at":"2026-06-05T04:43:18Z","source_collector":"spider:vendor_press","target_id":"https://arxiv.org/abs/2606.04560"},"tsa":{"authority":"crovia.substrate.bootstrap","rfc3161_token":"{\"kind\":\"crovia.bootstrap.tsa\",\"source_jsonl\":\"/opt/crovia/spider/data/news/vendor_press_v1.jsonl\",\"source_seal_merkle_root\":\"spider_vendor_press_v1\",\"upgrade_path\":\"Sessione H \\u2014 OpenTimestamps weekly anchor\"}"},"zk_mode":"clear","zk_proof":null},"ledger":{"leaf_hash":"d51423f4e5f4fd8a3a8ef6037aad35e4a484e4eb483e286e1d4f5ed71b3e6180","leaf_index":216050,"ledger_path":"/opt/crovia/substrate/axiom_ledger.jsonl"},"merkle_proof":{"hash_alg":"sha256","leaf_prefix":"0x00","node_prefix":"0x01","odd_leaf_rule":"duplicate_last","path":[{"sibling":"6ba854a4af06921848952c3224d2d4d63615526b3a5807502bb2cf4082034345","side":"right"},{"sibling":"84cdee7d9d3c8ceadaadc36a380d35a3f9d9329e8877317ce6f516ff2c325a06","side":"left"},{"sibling":"2518dcbb97417f5a8ea01e8bebabcc94bf89d1686d7ef6da23ceeb1b68d1efa4","side":"right"},{"sibling":"7db2dee1fbb6ef14733f4ef13a9243fe2b897a936ae26935d1d3e410aa6d8d22","side":"right"},{"sibling":"8bd2ae69e60460a30f0ed8f9b5a70d96bc4bd43656446e864a9d8c7f84b01b6d","side":"left"},{"sibling":"fa2fd112d39dac7cdc18ac1779f97d20ea942ce9f9a19774a22dcca910601984","side":"left"},{"sibling":"21ba3aa71af1cd7179c6bc067241197a4e092e6c68977def9c245081227eb701","side":"left"},{"sibling":"c232f0a98c6fd02eabc3155eb94d48261422543316c6f3198a92f09295fa8b65","side":"left"},{"sibling":"80c8f198b42e8ab9cecdef375223da498ae1c8691834ed79774638c44919af1a","side":"left"},{"sibling":"f71e5c82f0914ff67c825d0f6af89d5d4bac7dcf562dec6401a8c8ca4cb8b205","side":"left"},{"sibling":"16f9b88f45c6a3d5fd8e7a2c3900e7e76360c2a1a16cc1cd2ca83272e56ae03d","side":"right"},{"sibling":"37f2470d3b1c888c0179581620fded264990c2f86e80df0b5c379f1bd64cdb41","side":"left"},{"sibling":"24923c25478a136cad4a294856cb961d86e6a8aeb5b3c49e6e498039915fcf21","side":"right"},{"sibling":"10e59757ce3fa727404b5c4439d7171bdba5009288900a100ca9c9133d992bd6","side":"right"},{"sibling":"b2590791b920ca2a4ed39de126d2c0b1a10d9e7e62f572f12425f214e767b6e1","side":"left"},{"sibling":"084df6220452656a13c111887e1d2f105bb69fbb1e32c7e2ea96213f61503243","side":"right"},{"sibling":"c300cf0154c136afc09b1702a0be98f4ba5b6dc5cf57e8cc714ec1eaf4196eff","side":"left"},{"sibling":"d841ad93efda0869e5eb97678f348f03f5caab4353e05ff4bf18f47fb945b822","side":"left"}]},"schema":"crovia.axiom_proof.v1","seal":{"first_collector_run_id":"","first_receipt_hash":"","jsonl_path":"/opt/crovia/substrate/axiom_ledger.jsonl","key_id":"430895f101d38164","last_collector_run_id":"","last_receipt_hash":"","leaf_count":216172,"merkle_root":"459571be344a3b31e70215c839ba3e664ec8adfe4a34b6f4f0840892215ddbfd","public_key_hex":"cf742e26f75669dc673cb5c0786a1ae23ae8ca19c347317192ce40c28a7ff25c","run_id":"hourly_json_retrofit_20260605T053701Z","schema":"crovia.seal.v1","seal_family_version":"crovia-seal-family/1","seal_kind":"substrate_batch","sealed_at":"2026-06-05T05:38:28Z","sig_algorithm":"ed25519","signature":"d05ae4e1f30ece667b6ae35dce5276e69368ad7c979df791d19de6b9daf150d2ac5be528d75ff13fb5734608cc1c7b74afd6781d390d84517c87818d7d302d02","signer_version":"1.1.0"},"trust_root":{"key_id":"430895f101d38164","public_key_hex":"cf742e26f75669dc673cb5c0786a1ae23ae8ca19c347317192ce40c28a7ff25c","signature_algorithm":"ed25519","url":"/registry/canon/TRUST_ROOT.md"},"verifier":{"spec":"/registry/canon/AXIOM_RECEIPT_v1.md","url":"/v/axm_b1149fd71285f1c8af42d871df937e1ec6b6578cf0629be3d43b5298b8a49fc7"}}