{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:KOOOY6EIT5TYMEWJAVZE4RXDL3","short_pith_number":"pith:KOOOY6EI","schema_version":"1.0","canonical_sha256":"539cec78889f678612c905724e46e35ed6a2ac54bc41089a901355b0958d47c6","source":{"kind":"arxiv","id":"2605.14478","version":1},"attestation_state":"computed","paper":{"title":"When Retrieval Hurts Code Completion: A Diagnostic Study of Stale Repository Context","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Stale repository context actively biases code models toward generating outdated helper references.","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.SE","authors_text":"Haobin Pan, Hao Fu, Haojun Weng, Qianqian Yang, Xinwei Lv","submitted_at":"2026-05-14T07:18:30Z","abstract_excerpt":"Context: Retrieval-augmented code generation relies on cross-file repository context, but retrieved snippets may come from obsolete project states.\n  Objectives: We study whether temporally stale repository snippets act as harmless noise or actively induce current-state-incompatible code.\n  Methods: We conduct a controlled diagnostic study on a curated 17-sample set of production-helper signature changes from five Python repositories. For each sample, we compare current-only, stale-only, no-retrieval, and mixed current/stale retrieval conditions under prompts that hide commit freshness and exp"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2605.14478","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SE","submitted_at":"2026-05-14T07:18:30Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"4553a21d20eaa7c48991bf3a32b1010e027864b515607e39a4b2725a9729fd08","abstract_canon_sha256":"1464a870e7d605fb38c5b9105a1aad2e71bff0904458d47766ff1896f6d526bc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:39:06.578623Z","signature_b64":"ZYA3oF1cR6w+axuWSm5JwfcPTkWLDPTnktNoHx0v+EqQp3YbyCeA4YLNIuzwfNSqmCgBcqeNDCWzJEC25lC5BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"539cec78889f678612c905724e46e35ed6a2ac54bc41089a901355b0958d47c6","last_reissued_at":"2026-05-17T23:39:06.577950Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:39:06.577950Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"When Retrieval Hurts Code Completion: A Diagnostic Study of Stale Repository Context","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Stale repository context actively biases code models toward generating outdated helper references.","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.SE","authors_text":"Haobin Pan, Hao Fu, Haojun Weng, Qianqian Yang, Xinwei Lv","submitted_at":"2026-05-14T07:18:30Z","abstract_excerpt":"Context: Retrieval-augmented code generation relies on cross-file repository context, but retrieved snippets may come from obsolete project states.\n  Objectives: We study whether temporally stale repository snippets act as harmless noise or actively induce current-state-incompatible code.\n  Methods: We conduct a controlled diagnostic study on a curated 17-sample set of production-helper signature changes from five Python repositories. For each sample, we compare current-only, stale-only, no-retrieval, and mixed current/stale retrieval conditions under prompts that hide commit freshness and exp"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Under neutralized prompts, stale-only retrieval induces stale helper references on 15/17 Qwen2.5-Coder-7B-Instruct samples and 13/17 gpt-4.1-mini samples, corresponding to 88.2 and 76.5 percentage-point increases over current-only retrieval.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The curated 17-sample set of production-helper signature changes from five Python repositories is representative of typical real-world code completion scenarios and that the prompts successfully neutralize information about commit freshness.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Stale repository context in code RAG actively induces models to produce obsolete helper references, raising stale outputs by 76-88 percentage points over current-only retrieval in a 17-sample diagnostic study.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Stale repository context actively biases code models toward generating outdated helper references.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"3f1d616d502c8f165425ca0934a66ed3f471bf82d2ab34a4c2bfc76e1aadb590"},"source":{"id":"2605.14478","kind":"arxiv","version":1},"verdict":{"id":"d85206e5-54d3-45fc-b015-8926a9d7e432","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T02:02:26.828548Z","strongest_claim":"Under neutralized prompts, stale-only retrieval induces stale helper references on 15/17 Qwen2.5-Coder-7B-Instruct samples and 13/17 gpt-4.1-mini samples, corresponding to 88.2 and 76.5 percentage-point increases over current-only retrieval.","one_line_summary":"Stale repository context in code RAG actively induces models to produce obsolete helper references, raising stale outputs by 76-88 percentage points over current-only retrieval in a 17-sample diagnostic study.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The curated 17-sample set of production-helper signature changes from five Python repositories is representative of typical real-world code completion scenarios and that the prompts successfully neutralize information about commit freshness.","pith_extraction_headline":"Stale repository context actively biases code models toward generating outdated helper references."},"references":{"count":16,"sample":[{"doi":"","year":2026,"title":"When LLMs Lag Behind: Knowledge Conflicts from Evolving APIs in Code Generation","work_id":"3a725614-bcd0-44bc-befc-e31c9998d708","ref_index":1,"cited_arxiv_id":"2604.09515","is_internal_anchor":true},{"doi":"","year":2025,"title":"L. Liang, J. Gong, M. Liu, C. Wang, G. Ou, Y. Wang, X. Peng, Z. Zheng, RustEvo2: An Evolving Benchmark for API Evolution in LLM-based Rust Code Generation (2025).arXiv:2503.16922. URLhttps://arxiv.org","work_id":"df2cba79-8f28-4832-bdc6-cfccd9adc5cf","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"arXiv preprint arXiv:2303.12570 , year=","work_id":"a725c38c-a880-421f-bea1-b7fb1b5b105f","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"S. Zhang, Y. Ding, S. Lian, S. Song, H. Li, CodeRAG: Finding Relevant and Necessary Knowledge for Retrieval-Augmented Repository-Level Code Completion (2025).arXiv:2509.16112. URLhttps://arxiv.org/abs","work_id":"f61f1242-d5a2-42a5-b84f-6b872475a703","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"RepoBench: Benchmarking Repository-Level Code Auto-Completion Systems","work_id":"24b24ba5-b21c-43c6-866f-4b874305372e","ref_index":5,"cited_arxiv_id":"2306.03091","is_internal_anchor":true}],"resolved_work":16,"snapshot_sha256":"7f5200604086f0505d2b5079c25293601e30812be8907cd57cdf4405daf2e733","internal_anchors":3},"formal_canon":{"evidence_count":2,"snapshot_sha256":"1abc257c7539c1ef8f328adb29c1cc04b43bdb3459206d59195882c1bf0f646c"},"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":"2605.14478","created_at":"2026-05-17T23:39:06.578043+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.14478v1","created_at":"2026-05-17T23:39:06.578043+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.14478","created_at":"2026-05-17T23:39:06.578043+00:00"},{"alias_kind":"pith_short_12","alias_value":"KOOOY6EIT5TY","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"KOOOY6EIT5TYMEWJ","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"KOOOY6EI","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KOOOY6EIT5TYMEWJAVZE4RXDL3","json":"https://pith.science/pith/KOOOY6EIT5TYMEWJAVZE4RXDL3.json","graph_json":"https://pith.science/api/pith-number/KOOOY6EIT5TYMEWJAVZE4RXDL3/graph.json","events_json":"https://pith.science/api/pith-number/KOOOY6EIT5TYMEWJAVZE4RXDL3/events.json","paper":"https://pith.science/paper/KOOOY6EI"},"agent_actions":{"view_html":"https://pith.science/pith/KOOOY6EIT5TYMEWJAVZE4RXDL3","download_json":"https://pith.science/pith/KOOOY6EIT5TYMEWJAVZE4RXDL3.json","view_paper":"https://pith.science/paper/KOOOY6EI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.14478&json=true","fetch_graph":"https://pith.science/api/pith-number/KOOOY6EIT5TYMEWJAVZE4RXDL3/graph.json","fetch_events":"https://pith.science/api/pith-number/KOOOY6EIT5TYMEWJAVZE4RXDL3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KOOOY6EIT5TYMEWJAVZE4RXDL3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KOOOY6EIT5TYMEWJAVZE4RXDL3/action/storage_attestation","attest_author":"https://pith.science/pith/KOOOY6EIT5TYMEWJAVZE4RXDL3/action/author_attestation","sign_citation":"https://pith.science/pith/KOOOY6EIT5TYMEWJAVZE4RXDL3/action/citation_signature","submit_replication":"https://pith.science/pith/KOOOY6EIT5TYMEWJAVZE4RXDL3/action/replication_record"}},"created_at":"2026-05-17T23:39:06.578043+00:00","updated_at":"2026-05-17T23:39:06.578043+00:00"}