{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:7E45J63RBTWECNKSJB5O2K2YXR","short_pith_number":"pith:7E45J63R","schema_version":"1.0","canonical_sha256":"f939d4fb710cec413552487aed2b58bc73b8f5112332b8f9737a3c19fef5ca62","source":{"kind":"arxiv","id":"2410.24198","version":2},"attestation_state":"computed","paper":{"title":"SelfCodeAlign: Self-Alignment for Code Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","cs.SE"],"primary_cat":"cs.CL","authors_text":"Arjun Guha, Federico Cassano, Harm de Vries, Jiawei Liu, Leandro Von Werra, Lingming Zhang, Naman Jain, Yifeng Ding, Yuxiang Wei, Zachary Mueller","submitted_at":"2024-10-31T17:55:13Z","abstract_excerpt":"Instruction tuning is a supervised fine-tuning approach that significantly improves the ability of large language models (LLMs) to follow human instructions. We propose SelfCodeAlign, the first fully transparent and permissive pipeline for self-aligning code LLMs without extensive human annotations or distillation. SelfCodeAlign employs the same base model for inference throughout the data generation process. It first extracts diverse coding concepts from high-quality seed snippets to generate new tasks. It then samples multiple responses per task, pairs each with test cases, and validates the"},"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":"2410.24198","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-31T17:55:13Z","cross_cats_sorted":["cs.LG","cs.SE"],"title_canon_sha256":"451fc7f8c38f533b2ae4a4ef47d0e7d9d178f502e5c9524e6bee9ce8f537ce3b","abstract_canon_sha256":"6df5a48ffe49477c5aab60c25359310de7943b20b7a3b1fa722d646842347ced"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:29:36.166764Z","signature_b64":"0lRiZADLFJPpuNCGnXERxJY90YNWF+T8QdJw4YrhujGd+iPig+QNphYfj1QloblCBt3XDb79plFKM4h1FDsrDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f939d4fb710cec413552487aed2b58bc73b8f5112332b8f9737a3c19fef5ca62","last_reissued_at":"2026-07-05T09:29:36.166226Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:29:36.166226Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SelfCodeAlign: Self-Alignment for Code Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","cs.SE"],"primary_cat":"cs.CL","authors_text":"Arjun Guha, Federico Cassano, Harm de Vries, Jiawei Liu, Leandro Von Werra, Lingming Zhang, Naman Jain, Yifeng Ding, Yuxiang Wei, Zachary Mueller","submitted_at":"2024-10-31T17:55:13Z","abstract_excerpt":"Instruction tuning is a supervised fine-tuning approach that significantly improves the ability of large language models (LLMs) to follow human instructions. We propose SelfCodeAlign, the first fully transparent and permissive pipeline for self-aligning code LLMs without extensive human annotations or distillation. SelfCodeAlign employs the same base model for inference throughout the data generation process. It first extracts diverse coding concepts from high-quality seed snippets to generate new tasks. It then samples multiple responses per task, pairs each with test cases, and validates the"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.24198","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/2410.24198/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":"2410.24198","created_at":"2026-07-05T09:29:36.166309+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.24198v2","created_at":"2026-07-05T09:29:36.166309+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.24198","created_at":"2026-07-05T09:29:36.166309+00:00"},{"alias_kind":"pith_short_12","alias_value":"7E45J63RBTWE","created_at":"2026-07-05T09:29:36.166309+00:00"},{"alias_kind":"pith_short_16","alias_value":"7E45J63RBTWECNKS","created_at":"2026-07-05T09:29:36.166309+00:00"},{"alias_kind":"pith_short_8","alias_value":"7E45J63R","created_at":"2026-07-05T09:29:36.166309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.28998","citing_title":"Reward-Free Code Alignment from Pretrained or Fine-Tuned LLM: Unpacking the Trade-offs for Code Generation","ref_index":53,"is_internal_anchor":false},{"citing_arxiv_id":"2606.28438","citing_title":"When AI Reviews Its Own Code: Recursive Self-Training Collapse in Code LLMs","ref_index":164,"is_internal_anchor":false},{"citing_arxiv_id":"2502.02737","citing_title":"SmolLM2: When Smol Goes Big -- Data-Centric Training of a Small Language Model","ref_index":240,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22659","citing_title":"RealBench: A Repo-Level Code Generation Benchmark Aligned with Real-World Software Development Practices","ref_index":45,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7E45J63RBTWECNKSJB5O2K2YXR","json":"https://pith.science/pith/7E45J63RBTWECNKSJB5O2K2YXR.json","graph_json":"https://pith.science/api/pith-number/7E45J63RBTWECNKSJB5O2K2YXR/graph.json","events_json":"https://pith.science/api/pith-number/7E45J63RBTWECNKSJB5O2K2YXR/events.json","paper":"https://pith.science/paper/7E45J63R"},"agent_actions":{"view_html":"https://pith.science/pith/7E45J63RBTWECNKSJB5O2K2YXR","download_json":"https://pith.science/pith/7E45J63RBTWECNKSJB5O2K2YXR.json","view_paper":"https://pith.science/paper/7E45J63R","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.24198&json=true","fetch_graph":"https://pith.science/api/pith-number/7E45J63RBTWECNKSJB5O2K2YXR/graph.json","fetch_events":"https://pith.science/api/pith-number/7E45J63RBTWECNKSJB5O2K2YXR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7E45J63RBTWECNKSJB5O2K2YXR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7E45J63RBTWECNKSJB5O2K2YXR/action/storage_attestation","attest_author":"https://pith.science/pith/7E45J63RBTWECNKSJB5O2K2YXR/action/author_attestation","sign_citation":"https://pith.science/pith/7E45J63RBTWECNKSJB5O2K2YXR/action/citation_signature","submit_replication":"https://pith.science/pith/7E45J63RBTWECNKSJB5O2K2YXR/action/replication_record"}},"created_at":"2026-07-05T09:29:36.166309+00:00","updated_at":"2026-07-05T09:29:36.166309+00:00"}