{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:UZWMEAIEGKXOQHEET3XWJRTUWJ","short_pith_number":"pith:UZWMEAIE","schema_version":"1.0","canonical_sha256":"a66cc2010432aee81c849eef64c674b276fd0b5cfebbe99c49cffdb6768458a4","source":{"kind":"arxiv","id":"2404.03189","version":2},"attestation_state":"computed","paper":{"title":"The Probabilities Also Matter: A More Faithful Metric for Faithfulness of Free-Text Explanations in Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Maria Perez-Ortiz, Nicolas Heess, Noah Y. Siegel, Oana-Maria Camburu","submitted_at":"2024-04-04T04:20:04Z","abstract_excerpt":"In order to oversee advanced AI systems, it is important to understand their underlying decision-making process. When prompted, large language models (LLMs) can provide natural language explanations or reasoning traces that sound plausible and receive high ratings from human annotators. However, it is unclear to what extent these explanations are faithful, i.e., truly capture the factors responsible for the model's predictions. In this work, we introduce Correlational Explanatory Faithfulness (CEF), a metric that can be used in faithfulness tests based on input interventions. Previous metrics "},"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":"2404.03189","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-04T04:20:04Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c435a5e697dfdd25f1ad3b665ecaa44f70ca7a76c205a2d69aa42b758fbab285","abstract_canon_sha256":"a2cef688f466f63ac047cc9a6bad031599838e86f5fb1071a0872849aa2ceab3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:28:39.855321Z","signature_b64":"0+R8gpN4uWkuTyVa2fAEU3sXBBJuKVX08BeRwrmD34GhP+F0Bo9ikXyOKgMGgSKanrgbLJ8kgFtMGDuuRAM8Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a66cc2010432aee81c849eef64c674b276fd0b5cfebbe99c49cffdb6768458a4","last_reissued_at":"2026-07-05T08:28:39.854849Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:28:39.854849Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Probabilities Also Matter: A More Faithful Metric for Faithfulness of Free-Text Explanations in Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Maria Perez-Ortiz, Nicolas Heess, Noah Y. Siegel, Oana-Maria Camburu","submitted_at":"2024-04-04T04:20:04Z","abstract_excerpt":"In order to oversee advanced AI systems, it is important to understand their underlying decision-making process. When prompted, large language models (LLMs) can provide natural language explanations or reasoning traces that sound plausible and receive high ratings from human annotators. However, it is unclear to what extent these explanations are faithful, i.e., truly capture the factors responsible for the model's predictions. In this work, we introduce Correlational Explanatory Faithfulness (CEF), a metric that can be used in faithfulness tests based on input interventions. Previous metrics "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.03189","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/2404.03189/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":"2404.03189","created_at":"2026-07-05T08:28:39.854905+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.03189v2","created_at":"2026-07-05T08:28:39.854905+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.03189","created_at":"2026-07-05T08:28:39.854905+00:00"},{"alias_kind":"pith_short_12","alias_value":"UZWMEAIEGKXO","created_at":"2026-07-05T08:28:39.854905+00:00"},{"alias_kind":"pith_short_16","alias_value":"UZWMEAIEGKXOQHEE","created_at":"2026-07-05T08:28:39.854905+00:00"},{"alias_kind":"pith_short_8","alias_value":"UZWMEAIE","created_at":"2026-07-05T08:28:39.854905+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/UZWMEAIEGKXOQHEET3XWJRTUWJ","json":"https://pith.science/pith/UZWMEAIEGKXOQHEET3XWJRTUWJ.json","graph_json":"https://pith.science/api/pith-number/UZWMEAIEGKXOQHEET3XWJRTUWJ/graph.json","events_json":"https://pith.science/api/pith-number/UZWMEAIEGKXOQHEET3XWJRTUWJ/events.json","paper":"https://pith.science/paper/UZWMEAIE"},"agent_actions":{"view_html":"https://pith.science/pith/UZWMEAIEGKXOQHEET3XWJRTUWJ","download_json":"https://pith.science/pith/UZWMEAIEGKXOQHEET3XWJRTUWJ.json","view_paper":"https://pith.science/paper/UZWMEAIE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.03189&json=true","fetch_graph":"https://pith.science/api/pith-number/UZWMEAIEGKXOQHEET3XWJRTUWJ/graph.json","fetch_events":"https://pith.science/api/pith-number/UZWMEAIEGKXOQHEET3XWJRTUWJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UZWMEAIEGKXOQHEET3XWJRTUWJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UZWMEAIEGKXOQHEET3XWJRTUWJ/action/storage_attestation","attest_author":"https://pith.science/pith/UZWMEAIEGKXOQHEET3XWJRTUWJ/action/author_attestation","sign_citation":"https://pith.science/pith/UZWMEAIEGKXOQHEET3XWJRTUWJ/action/citation_signature","submit_replication":"https://pith.science/pith/UZWMEAIEGKXOQHEET3XWJRTUWJ/action/replication_record"}},"created_at":"2026-07-05T08:28:39.854905+00:00","updated_at":"2026-07-05T08:28:39.854905+00:00"}