{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:TNIOBM2GEET2PZLDD33LM6XWGX","short_pith_number":"pith:TNIOBM2G","schema_version":"1.0","canonical_sha256":"9b50e0b3462127a7e5631ef6b67af635e70375216ee601f19bc85dafd672dbf5","source":{"kind":"arxiv","id":"2502.01619","version":3},"attestation_state":"computed","paper":{"title":"Learning to Generate Unit Tests for Automated Debugging","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.SE","authors_text":"Archiki Prasad, Elias Stengel-Eskin, Justin Chih-Yao Chen, Mohit Bansal, Zaid Khan","submitted_at":"2025-02-03T18:51:43Z","abstract_excerpt":"Unit tests (UTs) play an instrumental role in assessing code correctness as well as providing feedback to large language models (LLMs), motivating automated test generation. However, we uncover a trade-off between generating unit test inputs that reveal errors when given a faulty code and correctly predicting the unit test output without access to the gold solution. To address this trade-off, we propose UTGen, which teaches LLMs to generate unit test inputs that reveal errors along with their correct expected outputs based on task descriptions. Since model-generated tests can provide noisy sig"},"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":"2502.01619","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SE","submitted_at":"2025-02-03T18:51:43Z","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"title_canon_sha256":"59e59f1a969b0377cb15d191e55838926cffd5ce50d5ff7b060517acf4e48284","abstract_canon_sha256":"0a2ea555f765116bd0a3ad324c5942b3a974bc692825e30b29dba64c66011bc3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:56:52.556212Z","signature_b64":"phGqpTpCKI98+0cAT/u+9ix77E1MIqwTyYgHWSCj1v0LElHD1LNkGsP8mdpMfOCyX84oyxkZOp9s5q88lB7WCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9b50e0b3462127a7e5631ef6b67af635e70375216ee601f19bc85dafd672dbf5","last_reissued_at":"2026-07-05T11:56:52.555750Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:56:52.555750Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Learning to Generate Unit Tests for Automated Debugging","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.SE","authors_text":"Archiki Prasad, Elias Stengel-Eskin, Justin Chih-Yao Chen, Mohit Bansal, Zaid Khan","submitted_at":"2025-02-03T18:51:43Z","abstract_excerpt":"Unit tests (UTs) play an instrumental role in assessing code correctness as well as providing feedback to large language models (LLMs), motivating automated test generation. However, we uncover a trade-off between generating unit test inputs that reveal errors when given a faulty code and correctly predicting the unit test output without access to the gold solution. To address this trade-off, we propose UTGen, which teaches LLMs to generate unit test inputs that reveal errors along with their correct expected outputs based on task descriptions. Since model-generated tests can provide noisy sig"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.01619","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/2502.01619/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":"2502.01619","created_at":"2026-07-05T11:56:52.555814+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.01619v3","created_at":"2026-07-05T11:56:52.555814+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.01619","created_at":"2026-07-05T11:56:52.555814+00:00"},{"alias_kind":"pith_short_12","alias_value":"TNIOBM2GEET2","created_at":"2026-07-05T11:56:52.555814+00:00"},{"alias_kind":"pith_short_16","alias_value":"TNIOBM2GEET2PZLD","created_at":"2026-07-05T11:56:52.555814+00:00"},{"alias_kind":"pith_short_8","alias_value":"TNIOBM2G","created_at":"2026-07-05T11:56:52.555814+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.05249","citing_title":"SWE-InfraBench: Evaluating Language Models on Cloud Infrastructure Code","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2604.25737","citing_title":"SAFEdit: Does Multi-Agent Decomposition Resolve the Reliability Challenges of Instructed Code Editing?","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07248","citing_title":"PaT: Planning-after-Trial for Efficient Test-Time Code Generation","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07864","citing_title":"ZeroCoder: Can LLMs Improve Code Generation Without Ground-Truth Supervision?","ref_index":31,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TNIOBM2GEET2PZLDD33LM6XWGX","json":"https://pith.science/pith/TNIOBM2GEET2PZLDD33LM6XWGX.json","graph_json":"https://pith.science/api/pith-number/TNIOBM2GEET2PZLDD33LM6XWGX/graph.json","events_json":"https://pith.science/api/pith-number/TNIOBM2GEET2PZLDD33LM6XWGX/events.json","paper":"https://pith.science/paper/TNIOBM2G"},"agent_actions":{"view_html":"https://pith.science/pith/TNIOBM2GEET2PZLDD33LM6XWGX","download_json":"https://pith.science/pith/TNIOBM2GEET2PZLDD33LM6XWGX.json","view_paper":"https://pith.science/paper/TNIOBM2G","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.01619&json=true","fetch_graph":"https://pith.science/api/pith-number/TNIOBM2GEET2PZLDD33LM6XWGX/graph.json","fetch_events":"https://pith.science/api/pith-number/TNIOBM2GEET2PZLDD33LM6XWGX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TNIOBM2GEET2PZLDD33LM6XWGX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TNIOBM2GEET2PZLDD33LM6XWGX/action/storage_attestation","attest_author":"https://pith.science/pith/TNIOBM2GEET2PZLDD33LM6XWGX/action/author_attestation","sign_citation":"https://pith.science/pith/TNIOBM2GEET2PZLDD33LM6XWGX/action/citation_signature","submit_replication":"https://pith.science/pith/TNIOBM2GEET2PZLDD33LM6XWGX/action/replication_record"}},"created_at":"2026-07-05T11:56:52.555814+00:00","updated_at":"2026-07-05T11:56:52.555814+00:00"}