{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:CDNMYRPUJX6WC5TYS6SEUWYE57","short_pith_number":"pith:CDNMYRPU","schema_version":"1.0","canonical_sha256":"10dacc45f44dfd61767897a44a5b04efc7b0b6e01b010d2904ec26f407a5d74b","source":{"kind":"arxiv","id":"2307.00897","version":3},"attestation_state":"computed","paper":{"title":"Fixing confirmation bias in feature attribution methods via semantic match","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"\\c{S}. \\.Ilker Birbil, Daniel Fernandez-Llaneza, Giovanni Cin\\`a, Iacer Calixto, Ludovico Deponte, Nishant Mishra, Rob Goedhart, Sandro Pezzelle, Tabea E. R\\\"ober","submitted_at":"2023-07-03T09:50:08Z","abstract_excerpt":"Feature attribution methods have become a staple method to disentangle the complex behavior of black box models. Despite their success, some scholars have argued that such methods suffer from a serious flaw: they do not allow a reliable interpretation in terms of human concepts. Simply put, visualizing an array of feature contributions is not enough for humans to conclude something about a model's internal representations, and confirmation bias can trick users into false beliefs about model behavior. We argue that a structured approach is required to test whether our hypotheses on the model ar"},"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":"2307.00897","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-07-03T09:50:08Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7465d27d60e5170c5c2e8734019f47ef85ce03466a7a911148a14f6e92dc6dfc","abstract_canon_sha256":"73bc30672704d3b4e896ba9add1f91a7c81fff8987bfacdda873dea4fd2235d1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:49:12.521195Z","signature_b64":"/pTV55sY+H4s0sQKN2tx4OD9g6vs7g/I8lAqZi4n9sRUOUQ0xPH9YQtAPV1GHDQaQPKXd8Pu1XVi4QQFGST7BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"10dacc45f44dfd61767897a44a5b04efc7b0b6e01b010d2904ec26f407a5d74b","last_reissued_at":"2026-07-05T07:49:12.520736Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:49:12.520736Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Fixing confirmation bias in feature attribution methods via semantic match","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"\\c{S}. \\.Ilker Birbil, Daniel Fernandez-Llaneza, Giovanni Cin\\`a, Iacer Calixto, Ludovico Deponte, Nishant Mishra, Rob Goedhart, Sandro Pezzelle, Tabea E. R\\\"ober","submitted_at":"2023-07-03T09:50:08Z","abstract_excerpt":"Feature attribution methods have become a staple method to disentangle the complex behavior of black box models. Despite their success, some scholars have argued that such methods suffer from a serious flaw: they do not allow a reliable interpretation in terms of human concepts. Simply put, visualizing an array of feature contributions is not enough for humans to conclude something about a model's internal representations, and confirmation bias can trick users into false beliefs about model behavior. We argue that a structured approach is required to test whether our hypotheses on the model ar"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2307.00897","kind":"arxiv","version":3},"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/2307.00897/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":"2307.00897","created_at":"2026-07-05T07:49:12.520797+00:00"},{"alias_kind":"arxiv_version","alias_value":"2307.00897v3","created_at":"2026-07-05T07:49:12.520797+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2307.00897","created_at":"2026-07-05T07:49:12.520797+00:00"},{"alias_kind":"pith_short_12","alias_value":"CDNMYRPUJX6W","created_at":"2026-07-05T07:49:12.520797+00:00"},{"alias_kind":"pith_short_16","alias_value":"CDNMYRPUJX6WC5TY","created_at":"2026-07-05T07:49:12.520797+00:00"},{"alias_kind":"pith_short_8","alias_value":"CDNMYRPU","created_at":"2026-07-05T07:49:12.520797+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.24619","citing_title":"Interpretable phenotyping of Heart Failure patients with Dutch discharge letters","ref_index":50,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CDNMYRPUJX6WC5TYS6SEUWYE57","json":"https://pith.science/pith/CDNMYRPUJX6WC5TYS6SEUWYE57.json","graph_json":"https://pith.science/api/pith-number/CDNMYRPUJX6WC5TYS6SEUWYE57/graph.json","events_json":"https://pith.science/api/pith-number/CDNMYRPUJX6WC5TYS6SEUWYE57/events.json","paper":"https://pith.science/paper/CDNMYRPU"},"agent_actions":{"view_html":"https://pith.science/pith/CDNMYRPUJX6WC5TYS6SEUWYE57","download_json":"https://pith.science/pith/CDNMYRPUJX6WC5TYS6SEUWYE57.json","view_paper":"https://pith.science/paper/CDNMYRPU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2307.00897&json=true","fetch_graph":"https://pith.science/api/pith-number/CDNMYRPUJX6WC5TYS6SEUWYE57/graph.json","fetch_events":"https://pith.science/api/pith-number/CDNMYRPUJX6WC5TYS6SEUWYE57/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CDNMYRPUJX6WC5TYS6SEUWYE57/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CDNMYRPUJX6WC5TYS6SEUWYE57/action/storage_attestation","attest_author":"https://pith.science/pith/CDNMYRPUJX6WC5TYS6SEUWYE57/action/author_attestation","sign_citation":"https://pith.science/pith/CDNMYRPUJX6WC5TYS6SEUWYE57/action/citation_signature","submit_replication":"https://pith.science/pith/CDNMYRPUJX6WC5TYS6SEUWYE57/action/replication_record"}},"created_at":"2026-07-05T07:49:12.520797+00:00","updated_at":"2026-07-05T07:49:12.520797+00:00"}