{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:QNVG5H2FFQGDAWTL3U7K3C4I44","short_pith_number":"pith:QNVG5H2F","schema_version":"1.0","canonical_sha256":"836a6e9f452c0c305a6bdd3ead8b88e73ad1e345e5ef37c088762cd38557d209","source":{"kind":"arxiv","id":"2411.05199","version":3},"attestation_state":"computed","paper":{"title":"CodeLutra: Boosting LLM Code Generation via Preference-Guided Refinement","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Leitian Tao, Ryan Rossi, Saayan Mitra, Tong Yu, Tung Mai, Xiang Chen, Yixuan Li","submitted_at":"2024-11-07T21:51:07Z","abstract_excerpt":"Large Language Models (LLMs) have revolutionized code generation but require significant resources and often over-generalize, limiting their task-specific efficiency. Fine-tuning smaller, open-source LLMs provides a cost-effective alternative. However, standard supervised approaches rely only on correct examples, missing valuable insights from failures. We introduce CodeLutra, a framework that leverages both correct and incorrect code attempts. Instead of using only correct solutions, CodeLutra applies iterative preference-based refinement, comparing successful and failed outputs to better app"},"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":"2411.05199","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-11-07T21:51:07Z","cross_cats_sorted":[],"title_canon_sha256":"6bc089edf503b54742c6f9683db17077b4fc1af2e84706ffc985d0a4cfd2e1d9","abstract_canon_sha256":"8fe6b8199e130510530fb13e751d30d78228a417c572ab4e32651405731eb035"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:27:21.442001Z","signature_b64":"3x3sfhKwLnZnbs3KcCXH/yIrnir577Z8ixTHPh8E25VG2TinuR3GjQxjtoaZU7aa/uhFPMFK1pEJLMVCLDsnDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"836a6e9f452c0c305a6bdd3ead8b88e73ad1e345e5ef37c088762cd38557d209","last_reissued_at":"2026-07-05T11:27:21.441560Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:27:21.441560Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CodeLutra: Boosting LLM Code Generation via Preference-Guided Refinement","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Leitian Tao, Ryan Rossi, Saayan Mitra, Tong Yu, Tung Mai, Xiang Chen, Yixuan Li","submitted_at":"2024-11-07T21:51:07Z","abstract_excerpt":"Large Language Models (LLMs) have revolutionized code generation but require significant resources and often over-generalize, limiting their task-specific efficiency. Fine-tuning smaller, open-source LLMs provides a cost-effective alternative. However, standard supervised approaches rely only on correct examples, missing valuable insights from failures. We introduce CodeLutra, a framework that leverages both correct and incorrect code attempts. Instead of using only correct solutions, CodeLutra applies iterative preference-based refinement, comparing successful and failed outputs to better app"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.05199","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/2411.05199/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":"2411.05199","created_at":"2026-07-05T11:27:21.441617+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.05199v3","created_at":"2026-07-05T11:27:21.441617+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.05199","created_at":"2026-07-05T11:27:21.441617+00:00"},{"alias_kind":"pith_short_12","alias_value":"QNVG5H2FFQGD","created_at":"2026-07-05T11:27:21.441617+00:00"},{"alias_kind":"pith_short_16","alias_value":"QNVG5H2FFQGDAWTL","created_at":"2026-07-05T11:27:21.441617+00:00"},{"alias_kind":"pith_short_8","alias_value":"QNVG5H2F","created_at":"2026-07-05T11:27:21.441617+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.21285","citing_title":"Curiosity by Design: An LLM-based Coding Assistant Asking Clarification Questions","ref_index":47,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QNVG5H2FFQGDAWTL3U7K3C4I44","json":"https://pith.science/pith/QNVG5H2FFQGDAWTL3U7K3C4I44.json","graph_json":"https://pith.science/api/pith-number/QNVG5H2FFQGDAWTL3U7K3C4I44/graph.json","events_json":"https://pith.science/api/pith-number/QNVG5H2FFQGDAWTL3U7K3C4I44/events.json","paper":"https://pith.science/paper/QNVG5H2F"},"agent_actions":{"view_html":"https://pith.science/pith/QNVG5H2FFQGDAWTL3U7K3C4I44","download_json":"https://pith.science/pith/QNVG5H2FFQGDAWTL3U7K3C4I44.json","view_paper":"https://pith.science/paper/QNVG5H2F","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.05199&json=true","fetch_graph":"https://pith.science/api/pith-number/QNVG5H2FFQGDAWTL3U7K3C4I44/graph.json","fetch_events":"https://pith.science/api/pith-number/QNVG5H2FFQGDAWTL3U7K3C4I44/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QNVG5H2FFQGDAWTL3U7K3C4I44/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QNVG5H2FFQGDAWTL3U7K3C4I44/action/storage_attestation","attest_author":"https://pith.science/pith/QNVG5H2FFQGDAWTL3U7K3C4I44/action/author_attestation","sign_citation":"https://pith.science/pith/QNVG5H2FFQGDAWTL3U7K3C4I44/action/citation_signature","submit_replication":"https://pith.science/pith/QNVG5H2FFQGDAWTL3U7K3C4I44/action/replication_record"}},"created_at":"2026-07-05T11:27:21.441617+00:00","updated_at":"2026-07-05T11:27:21.441617+00:00"}