{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:MG5AZAUAQDOLAFPGZTFVKHFW3F","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":"6faa90e04dfe741759bcd811d0d4ab079293ec9b8c4d1efbc04bef97dc5f8d0b","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2022-12-06T00:53:38Z","title_canon_sha256":"346334464ac04bbad7ca0f98c7bbcbadcfcf0aab5d04a090a151c92c4394b351"},"schema_version":"1.0","source":{"id":"2212.02684","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2212.02684","created_at":"2026-07-05T05:22:33Z"},{"alias_kind":"arxiv_version","alias_value":"2212.02684v1","created_at":"2026-07-05T05:22:33Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2212.02684","created_at":"2026-07-05T05:22:33Z"},{"alias_kind":"pith_short_12","alias_value":"MG5AZAUAQDOL","created_at":"2026-07-05T05:22:33Z"},{"alias_kind":"pith_short_16","alias_value":"MG5AZAUAQDOLAFPG","created_at":"2026-07-05T05:22:33Z"},{"alias_kind":"pith_short_8","alias_value":"MG5AZAUA","created_at":"2026-07-05T05:22:33Z"}],"graph_snapshots":[{"event_id":"sha256:d9625b569bd0019d8d080a57b84fbc46f38d2f897ecc8dc72a8653f5114f775f","target":"graph","created_at":"2026-07-05T05:22:33Z","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/2212.02684/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The Codex model has demonstrated extraordinary competence in synthesizing code from natural language problem descriptions. However, in order to reveal unknown failure modes and hidden biases, such large-scale models must be systematically subjected to multiple and diverse evaluation studies.\n  In this work, we evaluate the code synthesis capabilities of the Codex model based on a set of 115 Python problem statements from a popular competitive programming portal: HackerRank. Our evaluation shows that Codex is indeed proficient in Python, solving 96% of the problems in a zero-shot setting, and 1","authors_text":"Anjan Karmakar, Julian Aron Prenner, Marco D'Ambros, Romain Robbes","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2022-12-06T00:53:38Z","title":"Codex Hacks HackerRank: Memorization Issues and a Framework for Code Synthesis Evaluation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2212.02684","kind":"arxiv","version":1},"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":"sha256:4dafa85c8b8ffb36ad46d622df5488de3acdbe8ffe3a316ce5bba546c5980de8","target":"record","created_at":"2026-07-05T05:22:33Z","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":"6faa90e04dfe741759bcd811d0d4ab079293ec9b8c4d1efbc04bef97dc5f8d0b","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2022-12-06T00:53:38Z","title_canon_sha256":"346334464ac04bbad7ca0f98c7bbcbadcfcf0aab5d04a090a151c92c4394b351"},"schema_version":"1.0","source":{"id":"2212.02684","kind":"arxiv","version":1}},"canonical_sha256":"61ba0c828080dcb015e6cccb551cb6d964817b8140dcaa718e38ad836dd52be3","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"61ba0c828080dcb015e6cccb551cb6d964817b8140dcaa718e38ad836dd52be3","first_computed_at":"2026-07-05T05:22:33.342810Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T05:22:33.342810Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"k7FYrB+ocInYJONPmZSO7rnmF+oRQ3vRTpFrVseBMGC4nUIRB+Q6+lcKdKTe/SEkqSBEOhDH2st8MM3ht9ZQAQ==","signature_status":"signed_v1","signed_at":"2026-07-05T05:22:33.343148Z","signed_message":"canonical_sha256_bytes"},"source_id":"2212.02684","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4dafa85c8b8ffb36ad46d622df5488de3acdbe8ffe3a316ce5bba546c5980de8","sha256:d9625b569bd0019d8d080a57b84fbc46f38d2f897ecc8dc72a8653f5114f775f"],"state_sha256":"b07d119f48fe441799f2deb2048d2f8767b0811e3f771f729cc392fe8e38cd69"}