{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:5FHUUCHFCU6ICLIKWBOE7WEYAK","short_pith_number":"pith:5FHUUCHF","schema_version":"1.0","canonical_sha256":"e94f4a08e5153c812d0ab05c4fd89802890acf561a8cb2cefb3f2d3ec35b9e68","source":{"kind":"arxiv","id":"2504.09246","version":2},"attestation_state":"computed","paper":{"title":"Type-Constrained Code Generation with Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.PL"],"primary_cat":"cs.LG","authors_text":"Dawn Song, Hao Wang, Jingxuan He, Koushik Sen, Martin Vechev, Niels M\\\"undler","submitted_at":"2025-04-12T15:03:00Z","abstract_excerpt":"Large language models (LLMs) have achieved notable success in code generation. However, they still frequently produce uncompilable output because their next-token inference procedure does not model formal aspects of code. Although constrained decoding is a promising approach to alleviate this issue, it has only been applied to handle either domain-specific languages or syntactic features of general-purpose programming languages. However, LLMs frequently generate code with typing errors, which are beyond the domain of syntax and generally hard to adequately constrain. To address this challenge,"},"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":"2504.09246","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-04-12T15:03:00Z","cross_cats_sorted":["cs.PL"],"title_canon_sha256":"b7ad9244b3d534634021c8a5384c86fcbbaa32d4d55a0307403e0714abbeda06","abstract_canon_sha256":"6f568c08a95c7cde467a5818d0b2cc7089fbb695f6a48a474200c4bf9b82876a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:00:06.359441Z","signature_b64":"eobaJk413WuXkggLX8T4oYdLfivGft5c1t7vrdsDImj/I/Hq/NzHIPTxn0p+H/7JsA6f00BQQ0IN47JIvpIqAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e94f4a08e5153c812d0ab05c4fd89802890acf561a8cb2cefb3f2d3ec35b9e68","last_reissued_at":"2026-07-05T11:00:06.358939Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:00:06.358939Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Type-Constrained Code Generation with Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.PL"],"primary_cat":"cs.LG","authors_text":"Dawn Song, Hao Wang, Jingxuan He, Koushik Sen, Martin Vechev, Niels M\\\"undler","submitted_at":"2025-04-12T15:03:00Z","abstract_excerpt":"Large language models (LLMs) have achieved notable success in code generation. However, they still frequently produce uncompilable output because their next-token inference procedure does not model formal aspects of code. Although constrained decoding is a promising approach to alleviate this issue, it has only been applied to handle either domain-specific languages or syntactic features of general-purpose programming languages. However, LLMs frequently generate code with typing errors, which are beyond the domain of syntax and generally hard to adequately constrain. To address this challenge,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.09246","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/2504.09246/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":"2504.09246","created_at":"2026-07-05T11:00:06.358995+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.09246v2","created_at":"2026-07-05T11:00:06.358995+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.09246","created_at":"2026-07-05T11:00:06.358995+00:00"},{"alias_kind":"pith_short_12","alias_value":"5FHUUCHFCU6I","created_at":"2026-07-05T11:00:06.358995+00:00"},{"alias_kind":"pith_short_16","alias_value":"5FHUUCHFCU6ICLIK","created_at":"2026-07-05T11:00:06.358995+00:00"},{"alias_kind":"pith_short_8","alias_value":"5FHUUCHF","created_at":"2026-07-05T11:00:06.358995+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2602.06934","citing_title":"Implementing Grassroots Logic Programs with Multiagent Transition Systems and AI (Full Version)","ref_index":11,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5FHUUCHFCU6ICLIKWBOE7WEYAK","json":"https://pith.science/pith/5FHUUCHFCU6ICLIKWBOE7WEYAK.json","graph_json":"https://pith.science/api/pith-number/5FHUUCHFCU6ICLIKWBOE7WEYAK/graph.json","events_json":"https://pith.science/api/pith-number/5FHUUCHFCU6ICLIKWBOE7WEYAK/events.json","paper":"https://pith.science/paper/5FHUUCHF"},"agent_actions":{"view_html":"https://pith.science/pith/5FHUUCHFCU6ICLIKWBOE7WEYAK","download_json":"https://pith.science/pith/5FHUUCHFCU6ICLIKWBOE7WEYAK.json","view_paper":"https://pith.science/paper/5FHUUCHF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.09246&json=true","fetch_graph":"https://pith.science/api/pith-number/5FHUUCHFCU6ICLIKWBOE7WEYAK/graph.json","fetch_events":"https://pith.science/api/pith-number/5FHUUCHFCU6ICLIKWBOE7WEYAK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5FHUUCHFCU6ICLIKWBOE7WEYAK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5FHUUCHFCU6ICLIKWBOE7WEYAK/action/storage_attestation","attest_author":"https://pith.science/pith/5FHUUCHFCU6ICLIKWBOE7WEYAK/action/author_attestation","sign_citation":"https://pith.science/pith/5FHUUCHFCU6ICLIKWBOE7WEYAK/action/citation_signature","submit_replication":"https://pith.science/pith/5FHUUCHFCU6ICLIKWBOE7WEYAK/action/replication_record"}},"created_at":"2026-07-05T11:00:06.358995+00:00","updated_at":"2026-07-05T11:00:06.358995+00:00"}