{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:BIT5NWROT7FPQFCJVBFCPPTEXN","short_pith_number":"pith:BIT5NWRO","schema_version":"1.0","canonical_sha256":"0a27d6da2e9fcaf81449a84a27be64bb6e3e8541881900b0a23c77c455abf1b7","source":{"kind":"arxiv","id":"2503.02382","version":2},"attestation_state":"computed","paper":{"title":"An Efficient and Precise Training Data Construction Framework for Process-supervised Reward Model in Mathematical Reasoning","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Fuwei Cui, Jiajun Zhang, Qianlong Du, Wei Sun","submitted_at":"2025-03-04T08:18:46Z","abstract_excerpt":"Enhancing the mathematical reasoning capabilities of Large Language Models (LLMs) is of great scientific and practical significance. Researchers typically employ process-supervised reward models (PRMs) to guide the reasoning process, effectively improving the models' reasoning abilities. However, existing methods for constructing process supervision training data, such as manual annotation and per-step Monte Carlo estimation, are often costly or suffer from poor quality. To address these challenges, this paper introduces a framework called EpicPRM, which annotates each intermediate reasoning s"},"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":"2503.02382","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-03-04T08:18:46Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9dc1d201bb667db8acfc6a25783fd8b6e422447460282cf2db9f1b9018467198","abstract_canon_sha256":"1496746744ed260aa72cada00113fcb5e2c1d95ebd0b156c1e1f4fb32483195d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:41:39.791373Z","signature_b64":"I2h8RS9WPjtCsTkUU5h9KOWeiHZGasihs/e2wr8lRcW8/3o0TnspBKgsqL9WnLe7KThMhHnha6erx5/Y1QYFAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0a27d6da2e9fcaf81449a84a27be64bb6e3e8541881900b0a23c77c455abf1b7","last_reissued_at":"2026-07-05T11:41:39.790869Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:41:39.790869Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"An Efficient and Precise Training Data Construction Framework for Process-supervised Reward Model in Mathematical Reasoning","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Fuwei Cui, Jiajun Zhang, Qianlong Du, Wei Sun","submitted_at":"2025-03-04T08:18:46Z","abstract_excerpt":"Enhancing the mathematical reasoning capabilities of Large Language Models (LLMs) is of great scientific and practical significance. Researchers typically employ process-supervised reward models (PRMs) to guide the reasoning process, effectively improving the models' reasoning abilities. However, existing methods for constructing process supervision training data, such as manual annotation and per-step Monte Carlo estimation, are often costly or suffer from poor quality. To address these challenges, this paper introduces a framework called EpicPRM, which annotates each intermediate reasoning s"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.02382","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/2503.02382/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":"2503.02382","created_at":"2026-07-05T11:41:39.790922+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.02382v2","created_at":"2026-07-05T11:41:39.790922+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.02382","created_at":"2026-07-05T11:41:39.790922+00:00"},{"alias_kind":"pith_short_12","alias_value":"BIT5NWROT7FP","created_at":"2026-07-05T11:41:39.790922+00:00"},{"alias_kind":"pith_short_16","alias_value":"BIT5NWROT7FPQFCJ","created_at":"2026-07-05T11:41:39.790922+00:00"},{"alias_kind":"pith_short_8","alias_value":"BIT5NWRO","created_at":"2026-07-05T11:41:39.790922+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2510.10649","citing_title":"Unlocking Exploration in RLVR: Uncertainty-aware Advantage Shaping for Deeper Reasoning","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2604.02341","citing_title":"LLM Reasoning with Process Rewards for Outcome-Guided Steps","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17957","citing_title":"Process Reward Models Meet Planning: Generating Precise and Scalable Datasets for Step-Level Rewards","ref_index":13,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BIT5NWROT7FPQFCJVBFCPPTEXN","json":"https://pith.science/pith/BIT5NWROT7FPQFCJVBFCPPTEXN.json","graph_json":"https://pith.science/api/pith-number/BIT5NWROT7FPQFCJVBFCPPTEXN/graph.json","events_json":"https://pith.science/api/pith-number/BIT5NWROT7FPQFCJVBFCPPTEXN/events.json","paper":"https://pith.science/paper/BIT5NWRO"},"agent_actions":{"view_html":"https://pith.science/pith/BIT5NWROT7FPQFCJVBFCPPTEXN","download_json":"https://pith.science/pith/BIT5NWROT7FPQFCJVBFCPPTEXN.json","view_paper":"https://pith.science/paper/BIT5NWRO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.02382&json=true","fetch_graph":"https://pith.science/api/pith-number/BIT5NWROT7FPQFCJVBFCPPTEXN/graph.json","fetch_events":"https://pith.science/api/pith-number/BIT5NWROT7FPQFCJVBFCPPTEXN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BIT5NWROT7FPQFCJVBFCPPTEXN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BIT5NWROT7FPQFCJVBFCPPTEXN/action/storage_attestation","attest_author":"https://pith.science/pith/BIT5NWROT7FPQFCJVBFCPPTEXN/action/author_attestation","sign_citation":"https://pith.science/pith/BIT5NWROT7FPQFCJVBFCPPTEXN/action/citation_signature","submit_replication":"https://pith.science/pith/BIT5NWROT7FPQFCJVBFCPPTEXN/action/replication_record"}},"created_at":"2026-07-05T11:41:39.790922+00:00","updated_at":"2026-07-05T11:41:39.790922+00:00"}