{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:HQIXGBAVVXEEJQSBXATHJ7PVCC","short_pith_number":"pith:HQIXGBAV","schema_version":"1.0","canonical_sha256":"3c11730415adc844c241b82674fdf51082f6b97949327cda052434ec86512941","source":{"kind":"arxiv","id":"2205.08696","version":2},"attestation_state":"computed","paper":{"title":"The Solvability of Interpretability Evaluation Metrics","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Julie Shah, Yilun Zhou","submitted_at":"2022-05-18T02:52:03Z","abstract_excerpt":"Feature attribution methods are popular for explaining neural network predictions, and they are often evaluated on metrics such as comprehensiveness and sufficiency. In this paper, we highlight an intriguing property of these metrics: their solvability. Concretely, we can define the problem of optimizing an explanation for a metric, which can be solved by beam search. This observation leads to the obvious yet unaddressed question: why do we use explainers (e.g., LIME) not based on solving the target metric, if the metric value represents explanation quality? We present a series of investigatio"},"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":"2205.08696","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-05-18T02:52:03Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"9f4a45c240e9084ff8765f9b94d46a765ad9d728a80a1ca9566664169b383f2f","abstract_canon_sha256":"9b823354feaafbf408cf9b7bbd5bdc02dda45e57e19dee14c7cc5b11832eb671"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:38:27.939455Z","signature_b64":"KuDBTq2PaHiqHBvcLAs/W8crJSkKxe4/jSTgMEtBh+2A8600x7kRxfJw0fixLUXEFQaRAM6zqn0IDL8yz79aBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3c11730415adc844c241b82674fdf51082f6b97949327cda052434ec86512941","last_reissued_at":"2026-07-05T05:38:27.938863Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:38:27.938863Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Solvability of Interpretability Evaluation Metrics","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Julie Shah, Yilun Zhou","submitted_at":"2022-05-18T02:52:03Z","abstract_excerpt":"Feature attribution methods are popular for explaining neural network predictions, and they are often evaluated on metrics such as comprehensiveness and sufficiency. In this paper, we highlight an intriguing property of these metrics: their solvability. Concretely, we can define the problem of optimizing an explanation for a metric, which can be solved by beam search. This observation leads to the obvious yet unaddressed question: why do we use explainers (e.g., LIME) not based on solving the target metric, if the metric value represents explanation quality? We present a series of investigatio"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2205.08696","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/2205.08696/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":"2205.08696","created_at":"2026-07-05T05:38:27.938945+00:00"},{"alias_kind":"arxiv_version","alias_value":"2205.08696v2","created_at":"2026-07-05T05:38:27.938945+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2205.08696","created_at":"2026-07-05T05:38:27.938945+00:00"},{"alias_kind":"pith_short_12","alias_value":"HQIXGBAVVXEE","created_at":"2026-07-05T05:38:27.938945+00:00"},{"alias_kind":"pith_short_16","alias_value":"HQIXGBAVVXEEJQSB","created_at":"2026-07-05T05:38:27.938945+00:00"},{"alias_kind":"pith_short_8","alias_value":"HQIXGBAV","created_at":"2026-07-05T05:38:27.938945+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.08934","citing_title":"From Mechanistic to Compositional Interpretability","ref_index":68,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HQIXGBAVVXEEJQSBXATHJ7PVCC","json":"https://pith.science/pith/HQIXGBAVVXEEJQSBXATHJ7PVCC.json","graph_json":"https://pith.science/api/pith-number/HQIXGBAVVXEEJQSBXATHJ7PVCC/graph.json","events_json":"https://pith.science/api/pith-number/HQIXGBAVVXEEJQSBXATHJ7PVCC/events.json","paper":"https://pith.science/paper/HQIXGBAV"},"agent_actions":{"view_html":"https://pith.science/pith/HQIXGBAVVXEEJQSBXATHJ7PVCC","download_json":"https://pith.science/pith/HQIXGBAVVXEEJQSBXATHJ7PVCC.json","view_paper":"https://pith.science/paper/HQIXGBAV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2205.08696&json=true","fetch_graph":"https://pith.science/api/pith-number/HQIXGBAVVXEEJQSBXATHJ7PVCC/graph.json","fetch_events":"https://pith.science/api/pith-number/HQIXGBAVVXEEJQSBXATHJ7PVCC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HQIXGBAVVXEEJQSBXATHJ7PVCC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HQIXGBAVVXEEJQSBXATHJ7PVCC/action/storage_attestation","attest_author":"https://pith.science/pith/HQIXGBAVVXEEJQSBXATHJ7PVCC/action/author_attestation","sign_citation":"https://pith.science/pith/HQIXGBAVVXEEJQSBXATHJ7PVCC/action/citation_signature","submit_replication":"https://pith.science/pith/HQIXGBAVVXEEJQSBXATHJ7PVCC/action/replication_record"}},"created_at":"2026-07-05T05:38:27.938945+00:00","updated_at":"2026-07-05T05:38:27.938945+00:00"}