{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:J6HM76D5V7GU7UDVTZD6BUGJFQ","short_pith_number":"pith:J6HM76D5","schema_version":"1.0","canonical_sha256":"4f8ecff87dafcd4fd0759e47e0d0c92c1fe4a9095294a550dae62cb45c9d98e6","source":{"kind":"arxiv","id":"2211.10154","version":2},"attestation_state":"computed","paper":{"title":"CRAFT: Concept Recursive Activation FacTorization for Explainability","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Agustin Picard, David Vigouroux, Julien Colin, Louis Bethune, R\\'emi Cad\\`ene, Thibaut Boissin, Thomas Fel, Thomas Serre","submitted_at":"2022-11-17T14:22:47Z","abstract_excerpt":"Attribution methods, which employ heatmaps to identify the most influential regions of an image that impact model decisions, have gained widespread popularity as a type of explainability method. However, recent research has exposed the limited practical value of these methods, attributed in part to their narrow focus on the most prominent regions of an image -- revealing \"where\" the model looks, but failing to elucidate \"what\" the model sees in those areas. In this work, we try to fill in this gap with CRAFT -- a novel approach to identify both \"what\" and \"where\" by generating concept-based 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":"2211.10154","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2022-11-17T14:22:47Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"0985c5edd7c17a2a46898bc4dfc3f32fbc124d425459f9bd0d653d75f407803b","abstract_canon_sha256":"a94f660b12f033c5872060db9f55866171e2ebe971e949eb910b709111af7a82"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:56:01.177475Z","signature_b64":"GcheoXXhbQhjmBoWuqWc/OUtsof/5e7WvnhZaDKxOqTymIPH2Gwl6ktugFVrMsqne05mp7balufNP/JG20N8DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4f8ecff87dafcd4fd0759e47e0d0c92c1fe4a9095294a550dae62cb45c9d98e6","last_reissued_at":"2026-07-05T05:56:01.176962Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:56:01.176962Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CRAFT: Concept Recursive Activation FacTorization for Explainability","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Agustin Picard, David Vigouroux, Julien Colin, Louis Bethune, R\\'emi Cad\\`ene, Thibaut Boissin, Thomas Fel, Thomas Serre","submitted_at":"2022-11-17T14:22:47Z","abstract_excerpt":"Attribution methods, which employ heatmaps to identify the most influential regions of an image that impact model decisions, have gained widespread popularity as a type of explainability method. However, recent research has exposed the limited practical value of these methods, attributed in part to their narrow focus on the most prominent regions of an image -- revealing \"where\" the model looks, but failing to elucidate \"what\" the model sees in those areas. In this work, we try to fill in this gap with CRAFT -- a novel approach to identify both \"what\" and \"where\" by generating concept-based ex"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2211.10154","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/2211.10154/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":"2211.10154","created_at":"2026-07-05T05:56:01.177015+00:00"},{"alias_kind":"arxiv_version","alias_value":"2211.10154v2","created_at":"2026-07-05T05:56:01.177015+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2211.10154","created_at":"2026-07-05T05:56:01.177015+00:00"},{"alias_kind":"pith_short_12","alias_value":"J6HM76D5V7GU","created_at":"2026-07-05T05:56:01.177015+00:00"},{"alias_kind":"pith_short_16","alias_value":"J6HM76D5V7GU7UDV","created_at":"2026-07-05T05:56:01.177015+00:00"},{"alias_kind":"pith_short_8","alias_value":"J6HM76D5","created_at":"2026-07-05T05:56:01.177015+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.11203","citing_title":"FeatMap: Understanding image manipulation in the feature space and its implications for feature space geometry","ref_index":9,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/J6HM76D5V7GU7UDVTZD6BUGJFQ","json":"https://pith.science/pith/J6HM76D5V7GU7UDVTZD6BUGJFQ.json","graph_json":"https://pith.science/api/pith-number/J6HM76D5V7GU7UDVTZD6BUGJFQ/graph.json","events_json":"https://pith.science/api/pith-number/J6HM76D5V7GU7UDVTZD6BUGJFQ/events.json","paper":"https://pith.science/paper/J6HM76D5"},"agent_actions":{"view_html":"https://pith.science/pith/J6HM76D5V7GU7UDVTZD6BUGJFQ","download_json":"https://pith.science/pith/J6HM76D5V7GU7UDVTZD6BUGJFQ.json","view_paper":"https://pith.science/paper/J6HM76D5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2211.10154&json=true","fetch_graph":"https://pith.science/api/pith-number/J6HM76D5V7GU7UDVTZD6BUGJFQ/graph.json","fetch_events":"https://pith.science/api/pith-number/J6HM76D5V7GU7UDVTZD6BUGJFQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/J6HM76D5V7GU7UDVTZD6BUGJFQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/J6HM76D5V7GU7UDVTZD6BUGJFQ/action/storage_attestation","attest_author":"https://pith.science/pith/J6HM76D5V7GU7UDVTZD6BUGJFQ/action/author_attestation","sign_citation":"https://pith.science/pith/J6HM76D5V7GU7UDVTZD6BUGJFQ/action/citation_signature","submit_replication":"https://pith.science/pith/J6HM76D5V7GU7UDVTZD6BUGJFQ/action/replication_record"}},"created_at":"2026-07-05T05:56:01.177015+00:00","updated_at":"2026-07-05T05:56:01.177015+00:00"}