{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:YJDDTPZFAZKX44Q72EG6T4FH6A","short_pith_number":"pith:YJDDTPZF","schema_version":"1.0","canonical_sha256":"c24639bf2506557e721fd10de9f0a7f00160ec1bd91952ea5456bd06f81d3a3a","source":{"kind":"arxiv","id":"2410.02749","version":3},"attestation_state":"computed","paper":{"title":"Training Language Models on Synthetic Edit Sequences Improves Code Synthesis","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Lerrel Pinto, Rob Fergus, Ulyana Piterbarg","submitted_at":"2024-10-03T17:57:22Z","abstract_excerpt":"Software engineers mainly write code by editing existing programs. In contrast, language models (LMs) autoregressively synthesize programs in a single pass. One explanation for this is the scarcity of sequential edit data. While high-quality instruction data for code synthesis is scarce, edit data for synthesis is even scarcer. To fill this gap, we develop a synthetic data generation algorithm called LintSeq. This algorithm refactors programs into sequences of synthetic edits by using a linter to procedurally sample across interdependent lines of source code. Synthetic edits sampled with LintS"},"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":"2410.02749","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-03T17:57:22Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"a2d56698163528d7a86d55efa967e73cf1a3565e965e9b1198b42ca9e58d8093","abstract_canon_sha256":"4be45810e558507019cb88b6dd6cc3ef39e4621226fcb96524d58fed1f28eed2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:12:49.331294Z","signature_b64":"7+iQ2/xL4rTvTzGxXccckSA/6/QV2HYjm/M0IgQPL0JGsMJWdoop9BLnlDdwU995Ib2L+3wK3INLXLY5Ba6qBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c24639bf2506557e721fd10de9f0a7f00160ec1bd91952ea5456bd06f81d3a3a","last_reissued_at":"2026-07-05T10:12:49.330687Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:12:49.330687Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Training Language Models on Synthetic Edit Sequences Improves Code Synthesis","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Lerrel Pinto, Rob Fergus, Ulyana Piterbarg","submitted_at":"2024-10-03T17:57:22Z","abstract_excerpt":"Software engineers mainly write code by editing existing programs. In contrast, language models (LMs) autoregressively synthesize programs in a single pass. One explanation for this is the scarcity of sequential edit data. While high-quality instruction data for code synthesis is scarce, edit data for synthesis is even scarcer. To fill this gap, we develop a synthetic data generation algorithm called LintSeq. This algorithm refactors programs into sequences of synthetic edits by using a linter to procedurally sample across interdependent lines of source code. Synthetic edits sampled with LintS"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.02749","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/2410.02749/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":"2410.02749","created_at":"2026-07-05T10:12:49.330770+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.02749v3","created_at":"2026-07-05T10:12:49.330770+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.02749","created_at":"2026-07-05T10:12:49.330770+00:00"},{"alias_kind":"pith_short_12","alias_value":"YJDDTPZFAZKX","created_at":"2026-07-05T10:12:49.330770+00:00"},{"alias_kind":"pith_short_16","alias_value":"YJDDTPZFAZKX44Q7","created_at":"2026-07-05T10:12:49.330770+00:00"},{"alias_kind":"pith_short_8","alias_value":"YJDDTPZF","created_at":"2026-07-05T10:12:49.330770+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YJDDTPZFAZKX44Q72EG6T4FH6A","json":"https://pith.science/pith/YJDDTPZFAZKX44Q72EG6T4FH6A.json","graph_json":"https://pith.science/api/pith-number/YJDDTPZFAZKX44Q72EG6T4FH6A/graph.json","events_json":"https://pith.science/api/pith-number/YJDDTPZFAZKX44Q72EG6T4FH6A/events.json","paper":"https://pith.science/paper/YJDDTPZF"},"agent_actions":{"view_html":"https://pith.science/pith/YJDDTPZFAZKX44Q72EG6T4FH6A","download_json":"https://pith.science/pith/YJDDTPZFAZKX44Q72EG6T4FH6A.json","view_paper":"https://pith.science/paper/YJDDTPZF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.02749&json=true","fetch_graph":"https://pith.science/api/pith-number/YJDDTPZFAZKX44Q72EG6T4FH6A/graph.json","fetch_events":"https://pith.science/api/pith-number/YJDDTPZFAZKX44Q72EG6T4FH6A/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YJDDTPZFAZKX44Q72EG6T4FH6A/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YJDDTPZFAZKX44Q72EG6T4FH6A/action/storage_attestation","attest_author":"https://pith.science/pith/YJDDTPZFAZKX44Q72EG6T4FH6A/action/author_attestation","sign_citation":"https://pith.science/pith/YJDDTPZFAZKX44Q72EG6T4FH6A/action/citation_signature","submit_replication":"https://pith.science/pith/YJDDTPZFAZKX44Q72EG6T4FH6A/action/replication_record"}},"created_at":"2026-07-05T10:12:49.330770+00:00","updated_at":"2026-07-05T10:12:49.330770+00:00"}