{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:PMQVUH4WMCBMQQDHE3B2PAAZBO","short_pith_number":"pith:PMQVUH4W","schema_version":"1.0","canonical_sha256":"7b215a1f966082c8406726c3a780190b9261b32194a8e8527f182102f221c14b","source":{"kind":"arxiv","id":"2402.11456","version":2},"attestation_state":"computed","paper":{"title":"FactPICO: Factuality Evaluation for Plain Language Summarization of Medical Evidence","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Byron C Wallace, Hannah Louisa G\\\"oke, Jan Trienes, Junyi Jessy Li, Lily Chen, Monika Coers, Sebastian Antony Joseph, Wei Xu","submitted_at":"2024-02-18T04:45:01Z","abstract_excerpt":"Plain language summarization with LLMs can be useful for improving textual accessibility of technical content. But how factual are these summaries in a high-stakes domain like medicine? This paper presents FactPICO, a factuality benchmark for plain language summarization of medical texts describing randomized controlled trials (RCTs), which are the basis of evidence-based medicine and can directly inform patient treatment. FactPICO consists of 345 plain language summaries of RCT abstracts generated from three LLMs (i.e., GPT-4, Llama-2, and Alpaca), with fine-grained evaluation and natural lan"},"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":"2402.11456","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-18T04:45:01Z","cross_cats_sorted":[],"title_canon_sha256":"98583e613fd10d078ff35aa3ab7fcadc2ece3b1613d8b1b1b087f505a79eea85","abstract_canon_sha256":"56bb30e3a337ca9f4714b0454f013f94538855a3c6879b089db797920cc9fdf4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:27:24.813955Z","signature_b64":"AE2qaogOz75TGsNUR69QC4WjZSFPbD6YEOiLInrW53Ug/CvHaZ0+4pRkOU54A2yMpfo0eD0W0Z4GVPqeqqjsBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7b215a1f966082c8406726c3a780190b9261b32194a8e8527f182102f221c14b","last_reissued_at":"2026-07-05T08:27:24.813435Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:27:24.813435Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FactPICO: Factuality Evaluation for Plain Language Summarization of Medical Evidence","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Byron C Wallace, Hannah Louisa G\\\"oke, Jan Trienes, Junyi Jessy Li, Lily Chen, Monika Coers, Sebastian Antony Joseph, Wei Xu","submitted_at":"2024-02-18T04:45:01Z","abstract_excerpt":"Plain language summarization with LLMs can be useful for improving textual accessibility of technical content. But how factual are these summaries in a high-stakes domain like medicine? This paper presents FactPICO, a factuality benchmark for plain language summarization of medical texts describing randomized controlled trials (RCTs), which are the basis of evidence-based medicine and can directly inform patient treatment. FactPICO consists of 345 plain language summaries of RCT abstracts generated from three LLMs (i.e., GPT-4, Llama-2, and Alpaca), with fine-grained evaluation and natural lan"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.11456","kind":"arxiv","version":2},"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/2402.11456/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":"2402.11456","created_at":"2026-07-05T08:27:24.813493+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.11456v2","created_at":"2026-07-05T08:27:24.813493+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.11456","created_at":"2026-07-05T08:27:24.813493+00:00"},{"alias_kind":"pith_short_12","alias_value":"PMQVUH4WMCBM","created_at":"2026-07-05T08:27:24.813493+00:00"},{"alias_kind":"pith_short_16","alias_value":"PMQVUH4WMCBMQQDH","created_at":"2026-07-05T08:27:24.813493+00:00"},{"alias_kind":"pith_short_8","alias_value":"PMQVUH4W","created_at":"2026-07-05T08:27:24.813493+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/PMQVUH4WMCBMQQDHE3B2PAAZBO","json":"https://pith.science/pith/PMQVUH4WMCBMQQDHE3B2PAAZBO.json","graph_json":"https://pith.science/api/pith-number/PMQVUH4WMCBMQQDHE3B2PAAZBO/graph.json","events_json":"https://pith.science/api/pith-number/PMQVUH4WMCBMQQDHE3B2PAAZBO/events.json","paper":"https://pith.science/paper/PMQVUH4W"},"agent_actions":{"view_html":"https://pith.science/pith/PMQVUH4WMCBMQQDHE3B2PAAZBO","download_json":"https://pith.science/pith/PMQVUH4WMCBMQQDHE3B2PAAZBO.json","view_paper":"https://pith.science/paper/PMQVUH4W","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.11456&json=true","fetch_graph":"https://pith.science/api/pith-number/PMQVUH4WMCBMQQDHE3B2PAAZBO/graph.json","fetch_events":"https://pith.science/api/pith-number/PMQVUH4WMCBMQQDHE3B2PAAZBO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PMQVUH4WMCBMQQDHE3B2PAAZBO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PMQVUH4WMCBMQQDHE3B2PAAZBO/action/storage_attestation","attest_author":"https://pith.science/pith/PMQVUH4WMCBMQQDHE3B2PAAZBO/action/author_attestation","sign_citation":"https://pith.science/pith/PMQVUH4WMCBMQQDHE3B2PAAZBO/action/citation_signature","submit_replication":"https://pith.science/pith/PMQVUH4WMCBMQQDHE3B2PAAZBO/action/replication_record"}},"created_at":"2026-07-05T08:27:24.813493+00:00","updated_at":"2026-07-05T08:27:24.813493+00:00"}