{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:D4MEFV7NQQI72BPPYQRTD3F6WK","short_pith_number":"pith:D4MEFV7N","schema_version":"1.0","canonical_sha256":"1f1842d7ed8411fd05efc42331ecbeb28750bd129275a92d023e436f3af14a8d","source":{"kind":"arxiv","id":"2504.06122","version":3},"attestation_state":"computed","paper":{"title":"Leanabell-Prover: Posttraining Scaling in Formal Reasoning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Di Zhang, Fuzheng Zhang, Guorui Zhou, Jingyuan Zhang, Kun Gai, Qi Wang, Xingguang Ji, Yahui Liu, Yang Yue","submitted_at":"2025-04-08T15:15:26Z","abstract_excerpt":"Recent advances in automated theorem proving (ATP) through LLMs have highlighted the potential of formal reasoning with Lean 4 codes. However, ATP has not yet be revolutionized by the recent posttraining scaling as demonstrated by Open AI O1/O3 and Deepseek R1. In this work, we investigate the entire posttraining of ATP, aiming to align it with breakthroughs in reasoning models in natural languages. To begin, we continual train current ATP models with a hybrid dataset, which consists of numerous statement-proof pairs, and additional data aimed at incorporating cognitive behaviors that emulate "},"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":"2504.06122","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-04-08T15:15:26Z","cross_cats_sorted":[],"title_canon_sha256":"4a7abe5fca4e0d1daf6ceaad74c8fbc145fd2ade6cfb560c0962b250ba3f03d6","abstract_canon_sha256":"9e046881ca102dae703cfe1d379de5820dd2c5a9541b61dac3461d9aba44495a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:36:30.497427Z","signature_b64":"glhJ+n1/asWI2/0/EGrKVMl4w9xLN2MqmlJR6cXnhuN+zRbPg7lvwxN/gWCl3BHgeWzjfKme3w3MzkevBqXWDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1f1842d7ed8411fd05efc42331ecbeb28750bd129275a92d023e436f3af14a8d","last_reissued_at":"2026-07-05T11:36:30.496899Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:36:30.496899Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Leanabell-Prover: Posttraining Scaling in Formal Reasoning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Di Zhang, Fuzheng Zhang, Guorui Zhou, Jingyuan Zhang, Kun Gai, Qi Wang, Xingguang Ji, Yahui Liu, Yang Yue","submitted_at":"2025-04-08T15:15:26Z","abstract_excerpt":"Recent advances in automated theorem proving (ATP) through LLMs have highlighted the potential of formal reasoning with Lean 4 codes. However, ATP has not yet be revolutionized by the recent posttraining scaling as demonstrated by Open AI O1/O3 and Deepseek R1. In this work, we investigate the entire posttraining of ATP, aiming to align it with breakthroughs in reasoning models in natural languages. To begin, we continual train current ATP models with a hybrid dataset, which consists of numerous statement-proof pairs, and additional data aimed at incorporating cognitive behaviors that emulate "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.06122","kind":"arxiv","version":3},"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/2504.06122/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":"2504.06122","created_at":"2026-07-05T11:36:30.496948+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.06122v3","created_at":"2026-07-05T11:36:30.496948+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.06122","created_at":"2026-07-05T11:36:30.496948+00:00"},{"alias_kind":"pith_short_12","alias_value":"D4MEFV7NQQI7","created_at":"2026-07-05T11:36:30.496948+00:00"},{"alias_kind":"pith_short_16","alias_value":"D4MEFV7NQQI72BPP","created_at":"2026-07-05T11:36:30.496948+00:00"},{"alias_kind":"pith_short_8","alias_value":"D4MEFV7N","created_at":"2026-07-05T11:36:30.496948+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2607.07779","citing_title":"From Solvers to Research: Large Language Model-Driven Formal Mathematics at the Research Frontier","ref_index":295,"is_internal_anchor":true},{"citing_arxiv_id":"2607.05810","citing_title":"SCOPE: Leveraging Subgoal Critiques for Code Generation","ref_index":19,"is_internal_anchor":true},{"citing_arxiv_id":"2606.31134","citing_title":"Beyond the Library: An Agentic Framework for Autoformalizing Research Mathematics","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31134","citing_title":"Beyond the Library: An Agentic Framework for Autoformalizing Research Mathematics","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2601.13209","citing_title":"AI for Mathematics: Progress, Challenges, and Prospects","ref_index":172,"is_internal_anchor":false},{"citing_arxiv_id":"2510.01346","citing_title":"Aristotle: IMO-level Automated Theorem Proving","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19558","citing_title":"On Reasoning-Centric LLM-based Automated Theorem Proving","ref_index":31,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/D4MEFV7NQQI72BPPYQRTD3F6WK","json":"https://pith.science/pith/D4MEFV7NQQI72BPPYQRTD3F6WK.json","graph_json":"https://pith.science/api/pith-number/D4MEFV7NQQI72BPPYQRTD3F6WK/graph.json","events_json":"https://pith.science/api/pith-number/D4MEFV7NQQI72BPPYQRTD3F6WK/events.json","paper":"https://pith.science/paper/D4MEFV7N"},"agent_actions":{"view_html":"https://pith.science/pith/D4MEFV7NQQI72BPPYQRTD3F6WK","download_json":"https://pith.science/pith/D4MEFV7NQQI72BPPYQRTD3F6WK.json","view_paper":"https://pith.science/paper/D4MEFV7N","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.06122&json=true","fetch_graph":"https://pith.science/api/pith-number/D4MEFV7NQQI72BPPYQRTD3F6WK/graph.json","fetch_events":"https://pith.science/api/pith-number/D4MEFV7NQQI72BPPYQRTD3F6WK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/D4MEFV7NQQI72BPPYQRTD3F6WK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/D4MEFV7NQQI72BPPYQRTD3F6WK/action/storage_attestation","attest_author":"https://pith.science/pith/D4MEFV7NQQI72BPPYQRTD3F6WK/action/author_attestation","sign_citation":"https://pith.science/pith/D4MEFV7NQQI72BPPYQRTD3F6WK/action/citation_signature","submit_replication":"https://pith.science/pith/D4MEFV7NQQI72BPPYQRTD3F6WK/action/replication_record"}},"created_at":"2026-07-05T11:36:30.496948+00:00","updated_at":"2026-07-05T11:36:30.496948+00:00"}