{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:22ABGJOR47GY5DQ544C7N5I5AP","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":"bdc7ede44c0452bee59c77245df24f88dc9ca463c76d490d14c301f7b46657d1","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-24T17:59:05Z","title_canon_sha256":"73fa7dc1142c14c46879e3a1745145be2e46e4ce72183dd9c65330a8de8f39fd"},"schema_version":"1.0","source":{"id":"2407.17468","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2407.17468","created_at":"2026-07-05T08:48:08Z"},{"alias_kind":"arxiv_version","alias_value":"2407.17468v1","created_at":"2026-07-05T08:48:08Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.17468","created_at":"2026-07-05T08:48:08Z"},{"alias_kind":"pith_short_12","alias_value":"22ABGJOR47GY","created_at":"2026-07-05T08:48:08Z"},{"alias_kind":"pith_short_16","alias_value":"22ABGJOR47GY5DQ5","created_at":"2026-07-05T08:48:08Z"},{"alias_kind":"pith_short_8","alias_value":"22ABGJOR","created_at":"2026-07-05T08:48:08Z"}],"graph_snapshots":[{"event_id":"sha256:e911858b74c14905962cb3caa79cba6fc3c0272e44f104586e1812d72af53d84","target":"graph","created_at":"2026-07-05T08:48:08Z","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/2407.17468/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"While hallucinations of large language models (LLMs) prevail as a major challenge, existing evaluation benchmarks on factuality do not cover the diverse domains of knowledge that the real-world users of LLMs seek information about. To bridge this gap, we introduce WildHallucinations, a benchmark that evaluates factuality. It does so by prompting LLMs to generate information about entities mined from user-chatbot conversations in the wild. These generations are then automatically fact-checked against a systematically curated knowledge source collected from web search. Notably, half of these rea","authors_text":"Abhilasha Ravichander, Benjamin Newman, Claire Cardie, Khyathi Chandu, Liwei Jiang, Ronan Le Bras, Tanya Goyal, Wenting Zhao, Yejin Choi, Yuntian Deng, Yu Ying Chiu","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-24T17:59:05Z","title":"WildHallucinations: Evaluating Long-form Factuality in LLMs with Real-World Entity Queries"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.17468","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:f98fb29cad724da613d8f88146aab9fbc6964583243f4881c7b9e8bed28d7376","target":"record","created_at":"2026-07-05T08:48:08Z","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":"bdc7ede44c0452bee59c77245df24f88dc9ca463c76d490d14c301f7b46657d1","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-24T17:59:05Z","title_canon_sha256":"73fa7dc1142c14c46879e3a1745145be2e46e4ce72183dd9c65330a8de8f39fd"},"schema_version":"1.0","source":{"id":"2407.17468","kind":"arxiv","version":1}},"canonical_sha256":"d6801325d1e7cd8e8e1de705f6f51d03cf209184ee9732d1490b8c84c5388b0f","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"d6801325d1e7cd8e8e1de705f6f51d03cf209184ee9732d1490b8c84c5388b0f","first_computed_at":"2026-07-05T08:48:08.962931Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:48:08.962931Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Rs3fgqtiVn7dHP/csaFIeJ7NUMH0IaAEemCkpnGKWgjdkkhwADBIey/nH2GPgeJSQNuWQjJLmNbC4VUvMEDADA==","signature_status":"signed_v1","signed_at":"2026-07-05T08:48:08.963444Z","signed_message":"canonical_sha256_bytes"},"source_id":"2407.17468","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f98fb29cad724da613d8f88146aab9fbc6964583243f4881c7b9e8bed28d7376","sha256:e911858b74c14905962cb3caa79cba6fc3c0272e44f104586e1812d72af53d84"],"state_sha256":"342c2b689917652e6a76e7de45afd7c2c797984962478dc6bd8c1b243ffbcb39"}