{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:3NNJMMT73T6W7VQZHTTPUDQBI7","short_pith_number":"pith:3NNJMMT7","schema_version":"1.0","canonical_sha256":"db5a96327fdcfd6fd6193ce6fa0e0147f2ac3e27c5733f3c7416acf01c60404d","source":{"kind":"arxiv","id":"2501.01478","version":1},"attestation_state":"computed","paper":{"title":"Enhancing Reasoning through Process Supervision with Monte Carlo Tree Search","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.AI","authors_text":"Chaofan Ding, Kexin Luan, Shuaihao Dong, Shuangtao Li, Xinhan Di","submitted_at":"2025-01-02T12:09:17Z","abstract_excerpt":"Large language models (LLMs) have demonstrated their remarkable capacity across a variety of tasks. However, reasoning remains a challenge for LLMs. To improve LLMs' reasoning ability, process supervision has proven to be better than outcome supervision. In this work, we study using Monte Carlo Tree Search (MCTS) to generate process supervision data with LLMs themselves for training them. We sample reasoning steps with an LLM and assign each step a score that captures its \"relative correctness,\" and the LLM is then trained by minimizing weighted log-likelihood of generating the reasoning steps"},"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":"2501.01478","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2025-01-02T12:09:17Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"c54152fbd4b751de53ac669ddd236bbb29d7fca17cc20a77ee49908210d266fd","abstract_canon_sha256":"c9642113cada468c2d3fe2ef776efddc453cf9e99541ae2e995f8c4f1cf006ec"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:56:25.395376Z","signature_b64":"aa3IdXg/GqBTIZ+GzebJgoQ8yR9P9ZtfQuj6aMdS4y7lv7f9JUXkjuO4VmIyay8Dkg5PKBVuOXnA0dqAfQ5HDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"db5a96327fdcfd6fd6193ce6fa0e0147f2ac3e27c5733f3c7416acf01c60404d","last_reissued_at":"2026-07-05T09:56:25.394819Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:56:25.394819Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Enhancing Reasoning through Process Supervision with Monte Carlo Tree Search","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.AI","authors_text":"Chaofan Ding, Kexin Luan, Shuaihao Dong, Shuangtao Li, Xinhan Di","submitted_at":"2025-01-02T12:09:17Z","abstract_excerpt":"Large language models (LLMs) have demonstrated their remarkable capacity across a variety of tasks. However, reasoning remains a challenge for LLMs. To improve LLMs' reasoning ability, process supervision has proven to be better than outcome supervision. In this work, we study using Monte Carlo Tree Search (MCTS) to generate process supervision data with LLMs themselves for training them. We sample reasoning steps with an LLM and assign each step a score that captures its \"relative correctness,\" and the LLM is then trained by minimizing weighted log-likelihood of generating the reasoning steps"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.01478","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/2501.01478/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":"2501.01478","created_at":"2026-07-05T09:56:25.394891+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.01478v1","created_at":"2026-07-05T09:56:25.394891+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.01478","created_at":"2026-07-05T09:56:25.394891+00:00"},{"alias_kind":"pith_short_12","alias_value":"3NNJMMT73T6W","created_at":"2026-07-05T09:56:25.394891+00:00"},{"alias_kind":"pith_short_16","alias_value":"3NNJMMT73T6W7VQZ","created_at":"2026-07-05T09:56:25.394891+00:00"},{"alias_kind":"pith_short_8","alias_value":"3NNJMMT7","created_at":"2026-07-05T09:56:25.394891+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2502.00955","citing_title":"Efficient Multi-Agent System Training with Data Influence-Oriented Tree Search","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2502.11271","citing_title":"OctoTools: An Agentic Framework with Extensible Tools for Complex Reasoning","ref_index":44,"is_internal_anchor":false},{"citing_arxiv_id":"2507.17746","citing_title":"Rubrics as Rewards: Reinforcement Learning Beyond Verifiable Domains","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23333","citing_title":"Process Supervision of Confidence Margin for Calibrated LLM Reasoning","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2604.21268","citing_title":"Measure Twice, Click Once: Co-evolving Proposer and Visual Critic via Reinforcement Learning for GUI Grounding","ref_index":37,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3NNJMMT73T6W7VQZHTTPUDQBI7","json":"https://pith.science/pith/3NNJMMT73T6W7VQZHTTPUDQBI7.json","graph_json":"https://pith.science/api/pith-number/3NNJMMT73T6W7VQZHTTPUDQBI7/graph.json","events_json":"https://pith.science/api/pith-number/3NNJMMT73T6W7VQZHTTPUDQBI7/events.json","paper":"https://pith.science/paper/3NNJMMT7"},"agent_actions":{"view_html":"https://pith.science/pith/3NNJMMT73T6W7VQZHTTPUDQBI7","download_json":"https://pith.science/pith/3NNJMMT73T6W7VQZHTTPUDQBI7.json","view_paper":"https://pith.science/paper/3NNJMMT7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.01478&json=true","fetch_graph":"https://pith.science/api/pith-number/3NNJMMT73T6W7VQZHTTPUDQBI7/graph.json","fetch_events":"https://pith.science/api/pith-number/3NNJMMT73T6W7VQZHTTPUDQBI7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3NNJMMT73T6W7VQZHTTPUDQBI7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3NNJMMT73T6W7VQZHTTPUDQBI7/action/storage_attestation","attest_author":"https://pith.science/pith/3NNJMMT73T6W7VQZHTTPUDQBI7/action/author_attestation","sign_citation":"https://pith.science/pith/3NNJMMT73T6W7VQZHTTPUDQBI7/action/citation_signature","submit_replication":"https://pith.science/pith/3NNJMMT73T6W7VQZHTTPUDQBI7/action/replication_record"}},"created_at":"2026-07-05T09:56:25.394891+00:00","updated_at":"2026-07-05T09:56:25.394891+00:00"}