{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:HBIETVWSXR2ERP3IFOKMVZVGJ6","short_pith_number":"pith:HBIETVWS","schema_version":"1.0","canonical_sha256":"385049d6d2bc7448bf682b94cae6a64fa29c108409ace171d2c1c32da28b38c9","source":{"kind":"arxiv","id":"2503.15223","version":2},"attestation_state":"computed","paper":{"title":"Are \"Solved Issues\" in SWE-bench Really Solved Correctly? An Empirical Study","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.SE","authors_text":"Michael Pradel, You Wang, Zhongxin Liu","submitted_at":"2025-03-19T14:02:21Z","abstract_excerpt":"Automated issue solving aims to resolve real-world issues in software repositories. The most popular benchmarks for automated issue solving are SWE-bench and its human-filtered subset SWE-bench Verified. These benchmarks leverage testing to validate generated patches. However, because testing is rarely exhaustive, a patch may pass the tests but nevertheless fail to match the developers' expectations. Unfortunately, it is currently unclear to what extent evaluations performed with SWE-bench suffer from such plausible but incorrect patches. This paper presents an in-depth empirical study of 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":"2503.15223","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SE","submitted_at":"2025-03-19T14:02:21Z","cross_cats_sorted":[],"title_canon_sha256":"81a4e1ce3f7c2a94d4bc541d7cc371f5e33001671298526c5f9299e35cd6605a","abstract_canon_sha256":"2d5f4aa262e8840be3e47e47b6399c2a6588727ff327c06fd9f46e609e946550"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:07:04.077642Z","signature_b64":"3R1BvEkRwZX4KqDjbXBSWba9xi/g7ssI4aSttJ4r8Yr6mmhlzVUP1i03w1RaU4wyw4+g71Myv2U8Dt7hdeahCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"385049d6d2bc7448bf682b94cae6a64fa29c108409ace171d2c1c32da28b38c9","last_reissued_at":"2026-07-05T12:07:04.077184Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:07:04.077184Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Are \"Solved Issues\" in SWE-bench Really Solved Correctly? An Empirical Study","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.SE","authors_text":"Michael Pradel, You Wang, Zhongxin Liu","submitted_at":"2025-03-19T14:02:21Z","abstract_excerpt":"Automated issue solving aims to resolve real-world issues in software repositories. The most popular benchmarks for automated issue solving are SWE-bench and its human-filtered subset SWE-bench Verified. These benchmarks leverage testing to validate generated patches. However, because testing is rarely exhaustive, a patch may pass the tests but nevertheless fail to match the developers' expectations. Unfortunately, it is currently unclear to what extent evaluations performed with SWE-bench suffer from such plausible but incorrect patches. This paper presents an in-depth empirical study of the "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.15223","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/2503.15223/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":"2503.15223","created_at":"2026-07-05T12:07:04.077237+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.15223v2","created_at":"2026-07-05T12:07:04.077237+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.15223","created_at":"2026-07-05T12:07:04.077237+00:00"},{"alias_kind":"pith_short_12","alias_value":"HBIETVWSXR2E","created_at":"2026-07-05T12:07:04.077237+00:00"},{"alias_kind":"pith_short_16","alias_value":"HBIETVWSXR2ERP3I","created_at":"2026-07-05T12:07:04.077237+00:00"},{"alias_kind":"pith_short_8","alias_value":"HBIETVWS","created_at":"2026-07-05T12:07:04.077237+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":18,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2607.06184","citing_title":"What Resolve Rate Hides: Trajectory Structure Diagnostics for Coding Agents","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2607.06184","citing_title":"What Resolve Rate Hides: Trajectory Structure Diagnostics for Coding Agents","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2606.17799","citing_title":"Position: Coding Benchmarks Are Misaligned with Agentic Software Engineering","ref_index":46,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01360","citing_title":"Benchmarking Code Improvement with Progressive, Adaptive, and Interactive Feedback","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01360","citing_title":"Benchmarking Code Improvement with Progressive, Adaptive, and Interactive Feedback","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2606.28279","citing_title":"Agentic Hardware Design as Repository-Level Code Evolution","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2606.01326","citing_title":"Reducing Token Usage of State-in-Context Agents using Minification","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2606.05391","citing_title":"Human oversight of agentic systems in practice: Examining the oversight work, challenges, and heuristics of developers using software agents","ref_index":141,"is_internal_anchor":false},{"citing_arxiv_id":"2605.23262","citing_title":"Design and Report Benchmarks for Knowledge Work","ref_index":131,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21996","citing_title":"From Patches to Trajectories: Privileged Process Supervision for Software-Engineering Agents","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00505","citing_title":"LLM-Oriented Information Retrieval: A Denoising-First Perspective","ref_index":196,"is_internal_anchor":false},{"citing_arxiv_id":"2510.18270","citing_title":"Can Old Tests Do New Tricks for Resolving SWE Issues?","ref_index":40,"is_internal_anchor":false},{"citing_arxiv_id":"2601.18341","citing_title":"Agentic Much? Adoption of Coding Agents on GitHub","ref_index":40,"is_internal_anchor":false},{"citing_arxiv_id":"2602.18571","citing_title":"Debug2Fix: Can Interactive Debugging Help Coding Agents Fix More Bugs?","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10448","citing_title":"Can Agent Benchmarks Support Their Scores? Evidence-Supported Bounds for Interactive-Agent Evaluation","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00505","citing_title":"LLM-Oriented Information Retrieval: A Denoising-First Perspective","ref_index":189,"is_internal_anchor":false},{"citing_arxiv_id":"2604.04580","citing_title":"Beyond Fixed Tests: Repository-Level Issue Resolution as Coevolution of Code and Behavioral Constraints","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17351","citing_title":"SOCIA-EVO: Automated Simulator Construction via Dual-Anchored Bi-Level Optimization","ref_index":110,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HBIETVWSXR2ERP3IFOKMVZVGJ6","json":"https://pith.science/pith/HBIETVWSXR2ERP3IFOKMVZVGJ6.json","graph_json":"https://pith.science/api/pith-number/HBIETVWSXR2ERP3IFOKMVZVGJ6/graph.json","events_json":"https://pith.science/api/pith-number/HBIETVWSXR2ERP3IFOKMVZVGJ6/events.json","paper":"https://pith.science/paper/HBIETVWS"},"agent_actions":{"view_html":"https://pith.science/pith/HBIETVWSXR2ERP3IFOKMVZVGJ6","download_json":"https://pith.science/pith/HBIETVWSXR2ERP3IFOKMVZVGJ6.json","view_paper":"https://pith.science/paper/HBIETVWS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.15223&json=true","fetch_graph":"https://pith.science/api/pith-number/HBIETVWSXR2ERP3IFOKMVZVGJ6/graph.json","fetch_events":"https://pith.science/api/pith-number/HBIETVWSXR2ERP3IFOKMVZVGJ6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HBIETVWSXR2ERP3IFOKMVZVGJ6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HBIETVWSXR2ERP3IFOKMVZVGJ6/action/storage_attestation","attest_author":"https://pith.science/pith/HBIETVWSXR2ERP3IFOKMVZVGJ6/action/author_attestation","sign_citation":"https://pith.science/pith/HBIETVWSXR2ERP3IFOKMVZVGJ6/action/citation_signature","submit_replication":"https://pith.science/pith/HBIETVWSXR2ERP3IFOKMVZVGJ6/action/replication_record"}},"created_at":"2026-07-05T12:07:04.077237+00:00","updated_at":"2026-07-05T12:07:04.077237+00:00"}