{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:VBZJVNLXAMEHPPTNBZOQLMXD62","short_pith_number":"pith:VBZJVNLX","schema_version":"1.0","canonical_sha256":"a8729ab577030877be6d0e5d05b2e3f6b182e2a41085122b067e675a716c85ae","source":{"kind":"arxiv","id":"2405.07436","version":1},"attestation_state":"computed","paper":{"title":"Can Language Models Explain Their Own Classification Behavior?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Bilal Chughtai, Dane Sherburn, Owain Evans","submitted_at":"2024-05-13T02:31:08Z","abstract_excerpt":"Large language models (LLMs) perform well at a myriad of tasks, but explaining the processes behind this performance is a challenge. This paper investigates whether LLMs can give faithful high-level explanations of their own internal processes. To explore this, we introduce a dataset, ArticulateRules, of few-shot text-based classification tasks generated by simple rules. Each rule is associated with a simple natural-language explanation. We test whether models that have learned to classify inputs competently (both in- and out-of-distribution) are able to articulate freeform natural language ex"},"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":"2405.07436","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-05-13T02:31:08Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c3c94e9f6bfef8c08458d25369d38dde1ab3b5ed742fde71d1db23572c9b01ab","abstract_canon_sha256":"54df1d816637fa44e57f28f0bdefe99f2d7c224a023b4190ce459b540aa5b65b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:18:22.529173Z","signature_b64":"R0EbpIsILXmW+UVtNVNwm2wKSX9gfuU/debDtlAxKdu8k8B7ehEObIbhhk6q2t1MiqCjRlVGBE0BDS8sjcOIBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a8729ab577030877be6d0e5d05b2e3f6b182e2a41085122b067e675a716c85ae","last_reissued_at":"2026-07-05T08:18:22.528810Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:18:22.528810Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Can Language Models Explain Their Own Classification Behavior?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Bilal Chughtai, Dane Sherburn, Owain Evans","submitted_at":"2024-05-13T02:31:08Z","abstract_excerpt":"Large language models (LLMs) perform well at a myriad of tasks, but explaining the processes behind this performance is a challenge. This paper investigates whether LLMs can give faithful high-level explanations of their own internal processes. To explore this, we introduce a dataset, ArticulateRules, of few-shot text-based classification tasks generated by simple rules. Each rule is associated with a simple natural-language explanation. We test whether models that have learned to classify inputs competently (both in- and out-of-distribution) are able to articulate freeform natural language ex"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.07436","kind":"arxiv","version":1},"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/2405.07436/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":"2405.07436","created_at":"2026-07-05T08:18:22.528864+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.07436v1","created_at":"2026-07-05T08:18:22.528864+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.07436","created_at":"2026-07-05T08:18:22.528864+00:00"},{"alias_kind":"pith_short_12","alias_value":"VBZJVNLXAMEH","created_at":"2026-07-05T08:18:22.528864+00:00"},{"alias_kind":"pith_short_16","alias_value":"VBZJVNLXAMEHPPTN","created_at":"2026-07-05T08:18:22.528864+00:00"},{"alias_kind":"pith_short_8","alias_value":"VBZJVNLX","created_at":"2026-07-05T08:18:22.528864+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/VBZJVNLXAMEHPPTNBZOQLMXD62","json":"https://pith.science/pith/VBZJVNLXAMEHPPTNBZOQLMXD62.json","graph_json":"https://pith.science/api/pith-number/VBZJVNLXAMEHPPTNBZOQLMXD62/graph.json","events_json":"https://pith.science/api/pith-number/VBZJVNLXAMEHPPTNBZOQLMXD62/events.json","paper":"https://pith.science/paper/VBZJVNLX"},"agent_actions":{"view_html":"https://pith.science/pith/VBZJVNLXAMEHPPTNBZOQLMXD62","download_json":"https://pith.science/pith/VBZJVNLXAMEHPPTNBZOQLMXD62.json","view_paper":"https://pith.science/paper/VBZJVNLX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.07436&json=true","fetch_graph":"https://pith.science/api/pith-number/VBZJVNLXAMEHPPTNBZOQLMXD62/graph.json","fetch_events":"https://pith.science/api/pith-number/VBZJVNLXAMEHPPTNBZOQLMXD62/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VBZJVNLXAMEHPPTNBZOQLMXD62/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VBZJVNLXAMEHPPTNBZOQLMXD62/action/storage_attestation","attest_author":"https://pith.science/pith/VBZJVNLXAMEHPPTNBZOQLMXD62/action/author_attestation","sign_citation":"https://pith.science/pith/VBZJVNLXAMEHPPTNBZOQLMXD62/action/citation_signature","submit_replication":"https://pith.science/pith/VBZJVNLXAMEHPPTNBZOQLMXD62/action/replication_record"}},"created_at":"2026-07-05T08:18:22.528864+00:00","updated_at":"2026-07-05T08:18:22.528864+00:00"}