{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:3PSPJAU6N3QEXQJQPCH6OUVSN5","short_pith_number":"pith:3PSPJAU6","schema_version":"1.0","canonical_sha256":"dbe4f4829e6ee04bc130788fe752b26f45bac1c9162151a974f061ae6dc58222","source":{"kind":"arxiv","id":"2605.28022","version":1},"attestation_state":"computed","paper":{"title":"Beyond pass@k: Redundancy-Aware RLVR for Multi-Sample Code Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.SE"],"primary_cat":"cs.CL","authors_text":"Alexandre Verine, Benjamin Negrevergne, Le Bronnec Florian, Rio Yokota","submitted_at":"2026-05-27T06:26:52Z","abstract_excerpt":"LLMs for code generation are commonly evaluated in repeated-sampling settings using Pass@k, where multiple candidate programs are executed against unit tests under a finite sampling budget. While recent verifier-based reinforcement learning (RLVR) methods improve executable correctness, how these objectives affect redundancy among sampled programs remains poorly understood. In this work, we study implementation-level redundancy in code generation using JPlag, a plagiarism-detection system for code. Across models and benchmarks, we show that correctness-only RLVR often concentrates generations "},"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":"2605.28022","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-27T06:26:52Z","cross_cats_sorted":["cs.SE"],"title_canon_sha256":"8ef9da3855776f541f99330ef0a18d6b1bb37edcb408adee17d62d17d70d90d6","abstract_canon_sha256":"447e6529b108f23ad408db1296bed703332ab1fc8bb57f9e43aa91e6253cfda1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:04:56.250519Z","signature_b64":"G0Rc2S0WEgWZTwKrMLg8jOJFdHtYczdpoP10iQb9A025zeCQtTL1oF8IvhlB7oINbDacsC3muetJ70YvxkUdBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dbe4f4829e6ee04bc130788fe752b26f45bac1c9162151a974f061ae6dc58222","last_reissued_at":"2026-05-28T01:04:56.250064Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:04:56.250064Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Beyond pass@k: Redundancy-Aware RLVR for Multi-Sample Code Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.SE"],"primary_cat":"cs.CL","authors_text":"Alexandre Verine, Benjamin Negrevergne, Le Bronnec Florian, Rio Yokota","submitted_at":"2026-05-27T06:26:52Z","abstract_excerpt":"LLMs for code generation are commonly evaluated in repeated-sampling settings using Pass@k, where multiple candidate programs are executed against unit tests under a finite sampling budget. While recent verifier-based reinforcement learning (RLVR) methods improve executable correctness, how these objectives affect redundancy among sampled programs remains poorly understood. In this work, we study implementation-level redundancy in code generation using JPlag, a plagiarism-detection system for code. Across models and benchmarks, we show that correctness-only RLVR often concentrates generations "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.28022","kind":"arxiv","version":1},"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/2605.28022/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":"2605.28022","created_at":"2026-05-28T01:04:56.250115+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.28022v1","created_at":"2026-05-28T01:04:56.250115+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.28022","created_at":"2026-05-28T01:04:56.250115+00:00"},{"alias_kind":"pith_short_12","alias_value":"3PSPJAU6N3QE","created_at":"2026-05-28T01:04:56.250115+00:00"},{"alias_kind":"pith_short_16","alias_value":"3PSPJAU6N3QEXQJQ","created_at":"2026-05-28T01:04:56.250115+00:00"},{"alias_kind":"pith_short_8","alias_value":"3PSPJAU6","created_at":"2026-05-28T01:04:56.250115+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/3PSPJAU6N3QEXQJQPCH6OUVSN5","json":"https://pith.science/pith/3PSPJAU6N3QEXQJQPCH6OUVSN5.json","graph_json":"https://pith.science/api/pith-number/3PSPJAU6N3QEXQJQPCH6OUVSN5/graph.json","events_json":"https://pith.science/api/pith-number/3PSPJAU6N3QEXQJQPCH6OUVSN5/events.json","paper":"https://pith.science/paper/3PSPJAU6"},"agent_actions":{"view_html":"https://pith.science/pith/3PSPJAU6N3QEXQJQPCH6OUVSN5","download_json":"https://pith.science/pith/3PSPJAU6N3QEXQJQPCH6OUVSN5.json","view_paper":"https://pith.science/paper/3PSPJAU6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.28022&json=true","fetch_graph":"https://pith.science/api/pith-number/3PSPJAU6N3QEXQJQPCH6OUVSN5/graph.json","fetch_events":"https://pith.science/api/pith-number/3PSPJAU6N3QEXQJQPCH6OUVSN5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3PSPJAU6N3QEXQJQPCH6OUVSN5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3PSPJAU6N3QEXQJQPCH6OUVSN5/action/storage_attestation","attest_author":"https://pith.science/pith/3PSPJAU6N3QEXQJQPCH6OUVSN5/action/author_attestation","sign_citation":"https://pith.science/pith/3PSPJAU6N3QEXQJQPCH6OUVSN5/action/citation_signature","submit_replication":"https://pith.science/pith/3PSPJAU6N3QEXQJQPCH6OUVSN5/action/replication_record"}},"created_at":"2026-05-28T01:04:56.250115+00:00","updated_at":"2026-05-28T01:04:56.250115+00:00"}