{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:UB4WBFUXYW74OZ7P347OUR6NRT","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"ac0536302ecbfb854b627316bd84143d7251860885effc7e8bb13f08aba92dd5","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-12T15:21:36Z","title_canon_sha256":"77e7be7aca0145e3936733ba20e95963e24597f7014cc1bc21be5f837dbb7d81"},"schema_version":"1.0","source":{"id":"2401.06628","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2401.06628","created_at":"2026-07-05T07:47:28Z"},{"alias_kind":"arxiv_version","alias_value":"2401.06628v2","created_at":"2026-07-05T07:47:28Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.06628","created_at":"2026-07-05T07:47:28Z"},{"alias_kind":"pith_short_12","alias_value":"UB4WBFUXYW74","created_at":"2026-07-05T07:47:28Z"},{"alias_kind":"pith_short_16","alias_value":"UB4WBFUXYW74OZ7P","created_at":"2026-07-05T07:47:28Z"},{"alias_kind":"pith_short_8","alias_value":"UB4WBFUX","created_at":"2026-07-05T07:47:28Z"}],"graph_snapshots":[{"event_id":"sha256:c515acac7c87ac831d108c57cb4ffd1ddf47dcfce88420742ff752a4b9a3c80a","target":"graph","created_at":"2026-07-05T07:47:28Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2401.06628/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Advancing automated programming necessitates robust and comprehensive code generation benchmarks, yet current evaluation frameworks largely neglect object-oriented programming (OOP) in favor of functional programming (FP), e.g., HumanEval and MBPP. To address this, our study introduces a pioneering OOP-focused benchmark, featuring 431 Python programs that encompass essential OOP concepts and features like classes and encapsulation methods. We propose a novel evaluation metric, pass@o, tailored for OOP, enhancing traditional pass@k measures. Our evaluation of 23 leading large language models (L","authors_text":"Bo Du, Dacheng Tao, Liang Ding, Li Shen, Shuai Wang, Yong Luo","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-12T15:21:36Z","title":"OOP: Object-Oriented Programming Evaluation Benchmark for Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.06628","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:6a08d7da51f8b2ee52151b6014615d2ee68b3073336aa99721cb6b46b619e73b","target":"record","created_at":"2026-07-05T07:47:28Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"ac0536302ecbfb854b627316bd84143d7251860885effc7e8bb13f08aba92dd5","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-12T15:21:36Z","title_canon_sha256":"77e7be7aca0145e3936733ba20e95963e24597f7014cc1bc21be5f837dbb7d81"},"schema_version":"1.0","source":{"id":"2401.06628","kind":"arxiv","version":2}},"canonical_sha256":"a079609697c5bfc767efdf3eea47cd8cd8f45b7607f96e89d4166bfdcc904b9f","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a079609697c5bfc767efdf3eea47cd8cd8f45b7607f96e89d4166bfdcc904b9f","first_computed_at":"2026-07-05T07:47:28.042539Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:47:28.042539Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"twS2lX8E4Z4PArhZM4D8+T4dZb3HerjhN41Ju+jroyFJ0AJQ43lsymWo5LYt3Vf0Qc0oFg/0hIZ4FtXDs8TQAA==","signature_status":"signed_v1","signed_at":"2026-07-05T07:47:28.043115Z","signed_message":"canonical_sha256_bytes"},"source_id":"2401.06628","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6a08d7da51f8b2ee52151b6014615d2ee68b3073336aa99721cb6b46b619e73b","sha256:c515acac7c87ac831d108c57cb4ffd1ddf47dcfce88420742ff752a4b9a3c80a"],"state_sha256":"0bb9ff19efafa3aa2066c0dd1155cf24334e7434a73679a4c934a3610b63a6b0"}