{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:NVLQ2IZZF3HHL4EC5NFV3WQWVN","short_pith_number":"pith:NVLQ2IZZ","canonical_record":{"source":{"id":"2507.08267","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-11T02:26:01Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"bddc4cb7f5c6f8a1fb9691c3099e806c55be529f38019fd8a4739f7f282ca5aa","abstract_canon_sha256":"36280b93c40846fb0887a1e44aca7cd548b00fda09e3de0eff3e12471c6862fb"},"schema_version":"1.0"},"canonical_sha256":"6d570d23392ece75f082eb4b5dda16ab40f5c8d6bf033a864812b186bb4d82a0","source":{"kind":"arxiv","id":"2507.08267","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.08267","created_at":"2026-07-05T11:35:20Z"},{"alias_kind":"arxiv_version","alias_value":"2507.08267v1","created_at":"2026-07-05T11:35:20Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.08267","created_at":"2026-07-05T11:35:20Z"},{"alias_kind":"pith_short_12","alias_value":"NVLQ2IZZF3HH","created_at":"2026-07-05T11:35:20Z"},{"alias_kind":"pith_short_16","alias_value":"NVLQ2IZZF3HHL4EC","created_at":"2026-07-05T11:35:20Z"},{"alias_kind":"pith_short_8","alias_value":"NVLQ2IZZ","created_at":"2026-07-05T11:35:20Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:NVLQ2IZZF3HHL4EC5NFV3WQWVN","target":"record","payload":{"canonical_record":{"source":{"id":"2507.08267","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-11T02:26:01Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"bddc4cb7f5c6f8a1fb9691c3099e806c55be529f38019fd8a4739f7f282ca5aa","abstract_canon_sha256":"36280b93c40846fb0887a1e44aca7cd548b00fda09e3de0eff3e12471c6862fb"},"schema_version":"1.0"},"canonical_sha256":"6d570d23392ece75f082eb4b5dda16ab40f5c8d6bf033a864812b186bb4d82a0","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:35:20.361427Z","signature_b64":"r3hi5pno1eAcj0AhHQBpMycIY0XHq3M3fgh1qm6aR/tM3OrMF2TsbMomm8X8Gxy9CtNnAAElWpBVLgj0cPLfBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6d570d23392ece75f082eb4b5dda16ab40f5c8d6bf033a864812b186bb4d82a0","last_reissued_at":"2026-07-05T11:35:20.360835Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:35:20.360835Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2507.08267","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":"2026-07-05T11:35:20Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"MJt0KZ0qgDxID5FZrGgXADVAE3S55Fqvs0LpIllG4Tn3GEnBQGNMU7ElAe1ysv3dy5W7Z65TE4sI8V04RLimCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T04:57:37.787499Z"},"content_sha256":"dbdfb609ae77a4f0d35fe5a223c4d5b5778b3a0dbfc30d7cb908f8adf7007b5b","schema_version":"1.0","event_id":"sha256:dbdfb609ae77a4f0d35fe5a223c4d5b5778b3a0dbfc30d7cb908f8adf7007b5b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:NVLQ2IZZF3HHL4EC5NFV3WQWVN","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"A Practical Two-Stage Recipe for Mathematical LLMs: Maximizing Accuracy with SFT and Efficiency with Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Hiroshi Yoshihara, Taiki Yamaguchi, Yuichi Inoue","submitted_at":"2025-07-11T02:26:01Z","abstract_excerpt":"Enhancing the mathematical reasoning of Large Language Models (LLMs) is a pivotal challenge in advancing AI capabilities. While Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) are the dominant training paradigms, a systematic methodology for combining them to maximize both accuracy and efficiency remains largely unexplored. This paper introduces a practical and effective training recipe that strategically integrates extended SFT with RL from online inference (GRPO). We posit that these methods play complementary, not competing, roles: a prolonged SFT phase first pushes the model's"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.08267","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/2507.08267/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":"2026-07-05T11:35:20Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"VcrHUcEz/oIP9FegukEj/YbzMZRTLa6qdOa5CmLBjTARfJbmiWVjYELX+lmLkJ7s+9TNhYYzCjk4Qse2NPqkDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T04:57:37.788016Z"},"content_sha256":"167a85bb82a47efc585a88b7802c4f8a1bf83d7ed49d773f562f78965141feca","schema_version":"1.0","event_id":"sha256:167a85bb82a47efc585a88b7802c4f8a1bf83d7ed49d773f562f78965141feca"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/NVLQ2IZZF3HHL4EC5NFV3WQWVN/bundle.json","state_url":"https://pith.science/pith/NVLQ2IZZF3HHL4EC5NFV3WQWVN/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/NVLQ2IZZF3HHL4EC5NFV3WQWVN/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":"2026-08-09T04:57:37Z","links":{"resolver":"https://pith.science/pith/NVLQ2IZZF3HHL4EC5NFV3WQWVN","bundle":"https://pith.science/pith/NVLQ2IZZF3HHL4EC5NFV3WQWVN/bundle.json","state":"https://pith.science/pith/NVLQ2IZZF3HHL4EC5NFV3WQWVN/state.json","well_known_bundle":"https://pith.science/.well-known/pith/NVLQ2IZZF3HHL4EC5NFV3WQWVN/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:NVLQ2IZZF3HHL4EC5NFV3WQWVN","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":"36280b93c40846fb0887a1e44aca7cd548b00fda09e3de0eff3e12471c6862fb","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-11T02:26:01Z","title_canon_sha256":"bddc4cb7f5c6f8a1fb9691c3099e806c55be529f38019fd8a4739f7f282ca5aa"},"schema_version":"1.0","source":{"id":"2507.08267","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.08267","created_at":"2026-07-05T11:35:20Z"},{"alias_kind":"arxiv_version","alias_value":"2507.08267v1","created_at":"2026-07-05T11:35:20Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.08267","created_at":"2026-07-05T11:35:20Z"},{"alias_kind":"pith_short_12","alias_value":"NVLQ2IZZF3HH","created_at":"2026-07-05T11:35:20Z"},{"alias_kind":"pith_short_16","alias_value":"NVLQ2IZZF3HHL4EC","created_at":"2026-07-05T11:35:20Z"},{"alias_kind":"pith_short_8","alias_value":"NVLQ2IZZ","created_at":"2026-07-05T11:35:20Z"}],"graph_snapshots":[{"event_id":"sha256:167a85bb82a47efc585a88b7802c4f8a1bf83d7ed49d773f562f78965141feca","target":"graph","created_at":"2026-07-05T11:35:20Z","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/2507.08267/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Enhancing the mathematical reasoning of Large Language Models (LLMs) is a pivotal challenge in advancing AI capabilities. While Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) are the dominant training paradigms, a systematic methodology for combining them to maximize both accuracy and efficiency remains largely unexplored. This paper introduces a practical and effective training recipe that strategically integrates extended SFT with RL from online inference (GRPO). We posit that these methods play complementary, not competing, roles: a prolonged SFT phase first pushes the model's","authors_text":"Hiroshi Yoshihara, Taiki Yamaguchi, Yuichi Inoue","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-11T02:26:01Z","title":"A Practical Two-Stage Recipe for Mathematical LLMs: Maximizing Accuracy with SFT and Efficiency with Reinforcement Learning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.08267","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:dbdfb609ae77a4f0d35fe5a223c4d5b5778b3a0dbfc30d7cb908f8adf7007b5b","target":"record","created_at":"2026-07-05T11:35:20Z","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":"36280b93c40846fb0887a1e44aca7cd548b00fda09e3de0eff3e12471c6862fb","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-11T02:26:01Z","title_canon_sha256":"bddc4cb7f5c6f8a1fb9691c3099e806c55be529f38019fd8a4739f7f282ca5aa"},"schema_version":"1.0","source":{"id":"2507.08267","kind":"arxiv","version":1}},"canonical_sha256":"6d570d23392ece75f082eb4b5dda16ab40f5c8d6bf033a864812b186bb4d82a0","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"6d570d23392ece75f082eb4b5dda16ab40f5c8d6bf033a864812b186bb4d82a0","first_computed_at":"2026-07-05T11:35:20.360835Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:35:20.360835Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"r3hi5pno1eAcj0AhHQBpMycIY0XHq3M3fgh1qm6aR/tM3OrMF2TsbMomm8X8Gxy9CtNnAAElWpBVLgj0cPLfBw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:35:20.361427Z","signed_message":"canonical_sha256_bytes"},"source_id":"2507.08267","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:dbdfb609ae77a4f0d35fe5a223c4d5b5778b3a0dbfc30d7cb908f8adf7007b5b","sha256:167a85bb82a47efc585a88b7802c4f8a1bf83d7ed49d773f562f78965141feca"],"state_sha256":"60d25e1ab58ae12768fe84f534bfe29fe430782cad8815b2d991856306c87768"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"MJQAcd6XHeLQni+bG40Eg6V9GJLv4yLsxMDW/4R++sRJxLu0f7UbylhpNrV0zbsUbSfsu7nq3evcRuYixvT0Bg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T04:57:37.791694Z","bundle_sha256":"d7ff4a81893af56a00561d6bd78ea7be764e83d55c73ba520f4ba45d75c48672"}}