{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:XEX2BKIFHDDYFI2C375HL5IU53","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":"a8eb55f10cae05c2dac1a8d0d44fc017674f65af0d2209d3a5ef4798a8cdbc97","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-04-27T11:33:48Z","title_canon_sha256":"265026598bb536ddbe3166fef092e7e8d80ecb47fa826a0789a7b874d0beea94"},"schema_version":"1.0","source":{"id":"2304.14106","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2304.14106","created_at":"2026-07-05T08:32:59Z"},{"alias_kind":"arxiv_version","alias_value":"2304.14106v2","created_at":"2026-07-05T08:32:59Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2304.14106","created_at":"2026-07-05T08:32:59Z"},{"alias_kind":"pith_short_12","alias_value":"XEX2BKIFHDDY","created_at":"2026-07-05T08:32:59Z"},{"alias_kind":"pith_short_16","alias_value":"XEX2BKIFHDDYFI2C","created_at":"2026-07-05T08:32:59Z"},{"alias_kind":"pith_short_8","alias_value":"XEX2BKIF","created_at":"2026-07-05T08:32:59Z"}],"graph_snapshots":[{"event_id":"sha256:34b552b3b4a371ba2963b833a35171c30e0205f2c10c338acdd2b8c512f881a8","target":"graph","created_at":"2026-07-05T08:32:59Z","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/2304.14106/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"ChatGPT has achieved great success and can be considered to have acquired an infrastructural status. There are abundant works for evaluating ChatGPT on benchmarks. However, existing benchmarks encounter two challenges: (1) Disregard for periodical evaluation and (2) Lack of fine-grained features. In this paper, we construct ChatLog, an ever-updating dataset with large-scale records of diverse long-form ChatGPT responses for 21 NLP benchmarks from March, 2023 to now. We conduct a comprehensive performance evaluation to find that most capabilities of ChatGPT improve over time except for some abi","authors_text":"Chunyang Li, Jifan Yu, Juanzi Li, Lei Hou, Shangqing Tu, Xiaozhi Wang","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-04-27T11:33:48Z","title":"ChatLog: Carefully Evaluating the Evolution of ChatGPT Across Time"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2304.14106","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":"sha256:7d2c98fa9698f136766c398a8b2e5c680578d5f7e438d9e2968d0fc58e16403c","target":"record","created_at":"2026-07-05T08:32:59Z","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":"a8eb55f10cae05c2dac1a8d0d44fc017674f65af0d2209d3a5ef4798a8cdbc97","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-04-27T11:33:48Z","title_canon_sha256":"265026598bb536ddbe3166fef092e7e8d80ecb47fa826a0789a7b874d0beea94"},"schema_version":"1.0","source":{"id":"2304.14106","kind":"arxiv","version":2}},"canonical_sha256":"b92fa0a90538c782a342dffa75f514eee3479781693eb1e34bd296fae5f8ffc7","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"b92fa0a90538c782a342dffa75f514eee3479781693eb1e34bd296fae5f8ffc7","first_computed_at":"2026-07-05T08:32:59.053583Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:32:59.053583Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"u2qw4mVarJHSJsU0SXf2vk2kf2kVqlZbZymH/2j614ENxd0xIUvtKaOZC46XGXroc3S2weJdrwF9AL3hV2hdBA==","signature_status":"signed_v1","signed_at":"2026-07-05T08:32:59.054089Z","signed_message":"canonical_sha256_bytes"},"source_id":"2304.14106","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7d2c98fa9698f136766c398a8b2e5c680578d5f7e438d9e2968d0fc58e16403c","sha256:34b552b3b4a371ba2963b833a35171c30e0205f2c10c338acdd2b8c512f881a8"],"state_sha256":"32aeeffa01cf3bed7d5926bf70021f1b93919dd884912fa7691f5550f87b3d6a"}