{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:RICFXY3CW3T4OWDQZHBOURRE3L","short_pith_number":"pith:RICFXY3C","schema_version":"1.0","canonical_sha256":"8a045be362b6e7c75870c9c2ea4624daf47e86aa639ee9f3d4d8f991da1e606c","source":{"kind":"arxiv","id":"2105.06506","version":3},"attestation_state":"computed","paper":{"title":"Sanity Simulations for Saliency Methods","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Ameet Talwalkar, Gregory Plumb, Joon Sik Kim","submitted_at":"2021-05-13T18:40:57Z","abstract_excerpt":"Saliency methods are a popular class of feature attribution explanation methods that aim to capture a model's predictive reasoning by identifying \"important\" pixels in an input image. However, the development and adoption of these methods are hindered by the lack of access to ground-truth model reasoning, which prevents accurate evaluation. In this work, we design a synthetic benchmarking framework, SMERF, that allows us to perform ground-truth-based evaluation while controlling the complexity of the model's reasoning. Experimentally, SMERF reveals significant limitations in existing saliency "},"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":"2105.06506","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-05-13T18:40:57Z","cross_cats_sorted":[],"title_canon_sha256":"f38b4c1b4c12dd775ed1c0960d2f9a84654faa2d11ba3e8d337178bd1a0dfca5","abstract_canon_sha256":"68705aa1aa58999608286b9bebf1038c9181d41b0d357673badb7ad4b6725bc9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:32:25.831173Z","signature_b64":"Ulb3hUk/0JVWd8N9VAH329w4qL3TvWZrV2lybdtVXl3h+oJhSiHHS5q0pJAgF/4tpeGhxvwO2GLxnGSl9KviCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8a045be362b6e7c75870c9c2ea4624daf47e86aa639ee9f3d4d8f991da1e606c","last_reissued_at":"2026-07-05T04:32:25.830699Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:32:25.830699Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Sanity Simulations for Saliency Methods","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Ameet Talwalkar, Gregory Plumb, Joon Sik Kim","submitted_at":"2021-05-13T18:40:57Z","abstract_excerpt":"Saliency methods are a popular class of feature attribution explanation methods that aim to capture a model's predictive reasoning by identifying \"important\" pixels in an input image. However, the development and adoption of these methods are hindered by the lack of access to ground-truth model reasoning, which prevents accurate evaluation. In this work, we design a synthetic benchmarking framework, SMERF, that allows us to perform ground-truth-based evaluation while controlling the complexity of the model's reasoning. Experimentally, SMERF reveals significant limitations in existing saliency "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2105.06506","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/2105.06506/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":"2105.06506","created_at":"2026-07-05T04:32:25.830785+00:00"},{"alias_kind":"arxiv_version","alias_value":"2105.06506v3","created_at":"2026-07-05T04:32:25.830785+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2105.06506","created_at":"2026-07-05T04:32:25.830785+00:00"},{"alias_kind":"pith_short_12","alias_value":"RICFXY3CW3T4","created_at":"2026-07-05T04:32:25.830785+00:00"},{"alias_kind":"pith_short_16","alias_value":"RICFXY3CW3T4OWDQ","created_at":"2026-07-05T04:32:25.830785+00:00"},{"alias_kind":"pith_short_8","alias_value":"RICFXY3C","created_at":"2026-07-05T04:32:25.830785+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.11502","citing_title":"AIM: Amending Inherent Interpretability via Self-Supervised Masking","ref_index":50,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RICFXY3CW3T4OWDQZHBOURRE3L","json":"https://pith.science/pith/RICFXY3CW3T4OWDQZHBOURRE3L.json","graph_json":"https://pith.science/api/pith-number/RICFXY3CW3T4OWDQZHBOURRE3L/graph.json","events_json":"https://pith.science/api/pith-number/RICFXY3CW3T4OWDQZHBOURRE3L/events.json","paper":"https://pith.science/paper/RICFXY3C"},"agent_actions":{"view_html":"https://pith.science/pith/RICFXY3CW3T4OWDQZHBOURRE3L","download_json":"https://pith.science/pith/RICFXY3CW3T4OWDQZHBOURRE3L.json","view_paper":"https://pith.science/paper/RICFXY3C","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2105.06506&json=true","fetch_graph":"https://pith.science/api/pith-number/RICFXY3CW3T4OWDQZHBOURRE3L/graph.json","fetch_events":"https://pith.science/api/pith-number/RICFXY3CW3T4OWDQZHBOURRE3L/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RICFXY3CW3T4OWDQZHBOURRE3L/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RICFXY3CW3T4OWDQZHBOURRE3L/action/storage_attestation","attest_author":"https://pith.science/pith/RICFXY3CW3T4OWDQZHBOURRE3L/action/author_attestation","sign_citation":"https://pith.science/pith/RICFXY3CW3T4OWDQZHBOURRE3L/action/citation_signature","submit_replication":"https://pith.science/pith/RICFXY3CW3T4OWDQZHBOURRE3L/action/replication_record"}},"created_at":"2026-07-05T04:32:25.830785+00:00","updated_at":"2026-07-05T04:32:25.830785+00:00"}