{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:7FPBTNJ4FCKFPH4GC6L3YIGV3N","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":"30387b114f16b9dac5399779bd1652acf81d1043cad2c965f33f4f36f1d80d8f","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-07-17T17:41:47Z","title_canon_sha256":"f80ed4c98dc200b01d5b53d71892832503de26449a434845185ab355a5380812"},"schema_version":"1.0","source":{"id":"2307.08678","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2307.08678","created_at":"2026-07-05T06:31:27Z"},{"alias_kind":"arxiv_version","alias_value":"2307.08678v1","created_at":"2026-07-05T06:31:27Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2307.08678","created_at":"2026-07-05T06:31:27Z"},{"alias_kind":"pith_short_12","alias_value":"7FPBTNJ4FCKF","created_at":"2026-07-05T06:31:27Z"},{"alias_kind":"pith_short_16","alias_value":"7FPBTNJ4FCKFPH4G","created_at":"2026-07-05T06:31:27Z"},{"alias_kind":"pith_short_8","alias_value":"7FPBTNJ4","created_at":"2026-07-05T06:31:27Z"}],"graph_snapshots":[{"event_id":"sha256:04d5520f67fefb05355bcf2120ae09a04d97baae76a3c7d1fd0f23f7c8906d79","target":"graph","created_at":"2026-07-05T06:31:27Z","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/2307.08678/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) are trained to imitate humans to explain human decisions. However, do LLMs explain themselves? Can they help humans build mental models of how LLMs process different inputs? To answer these questions, we propose to evaluate $\\textbf{counterfactual simulatability}$ of natural language explanations: whether an explanation can enable humans to precisely infer the model's outputs on diverse counterfactuals of the explained input. For example, if a model answers \"yes\" to the input question \"Can eagles fly?\" with the explanation \"all birds can fly\", then humans would inf","authors_text":"Chen Zhao, He He, Jacob Steinhardt, Kathleen McKeown, Narutatsu Ri, Ruiqi Zhong, Yanda Chen, Zhou Yu","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-07-17T17:41:47Z","title":"Do Models Explain Themselves? Counterfactual Simulatability of Natural Language Explanations"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2307.08678","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":"sha256:6c882d1d4ce6adee0d0fbe63eda90aa8b0e33247027b4b9b45a0d92ee6905800","target":"record","created_at":"2026-07-05T06:31:27Z","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":"30387b114f16b9dac5399779bd1652acf81d1043cad2c965f33f4f36f1d80d8f","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-07-17T17:41:47Z","title_canon_sha256":"f80ed4c98dc200b01d5b53d71892832503de26449a434845185ab355a5380812"},"schema_version":"1.0","source":{"id":"2307.08678","kind":"arxiv","version":1}},"canonical_sha256":"f95e19b53c2894579f861797bc20d5db5dda4c02092fd34090a668fa45426e8d","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f95e19b53c2894579f861797bc20d5db5dda4c02092fd34090a668fa45426e8d","first_computed_at":"2026-07-05T06:31:27.545757Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T06:31:27.545757Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"wuy3Xwd1DDZI7y5o8sJ2AitrlhRayA+nIUQvG44kja3P0a3I9lzeO+5ehsORhj3Vinpt2K0K8vUJxoQQFmE6DQ==","signature_status":"signed_v1","signed_at":"2026-07-05T06:31:27.546361Z","signed_message":"canonical_sha256_bytes"},"source_id":"2307.08678","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6c882d1d4ce6adee0d0fbe63eda90aa8b0e33247027b4b9b45a0d92ee6905800","sha256:04d5520f67fefb05355bcf2120ae09a04d97baae76a3c7d1fd0f23f7c8906d79"],"state_sha256":"86d00ef2ede324c0bd1d2c0df9f40d530d82aa6e4f67afa216cc7301e8acda58"}