{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:R444ZXKMH7HUY4PNT2LQQRE2A7","short_pith_number":"pith:R444ZXKM","schema_version":"1.0","canonical_sha256":"8f39ccdd4c3fcf4c71ed9e9708449a07fe575cb85043f2ee90ad399e7bc72f25","source":{"kind":"arxiv","id":"2307.10563","version":1},"attestation_state":"computed","paper":{"title":"FACADE: A Framework for Adversarial Circuit Anomaly Detection and Evaluation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Andres Carranza, Arnuv Tandon, Dhruv Pai, Rylan Schaeffer, Sanmi Koyejo","submitted_at":"2023-07-20T04:00:37Z","abstract_excerpt":"We present FACADE, a novel probabilistic and geometric framework designed for unsupervised mechanistic anomaly detection in deep neural networks. Its primary goal is advancing the understanding and mitigation of adversarial attacks. FACADE aims to generate probabilistic distributions over circuits, which provide critical insights to their contribution to changes in the manifold properties of pseudo-classes, or high-dimensional modes in activation space, yielding a powerful tool for uncovering and combating adversarial attacks. Our approach seeks to improve model robustness, enhance scalable mo"},"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":"2307.10563","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-07-20T04:00:37Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"059503b0e59ca06f41b06fcdf5db9f275665a3ab21b91f93fc19653ac2044303","abstract_canon_sha256":"2c8ee71ef974f33ae131b7e1cb50514b2a2a99d8ed295811baf0b93e5de175db"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:33:07.084633Z","signature_b64":"ZgNjlBCT/pYHJbUCcsOhcmYbbLn8DSJ3XaH3IlrqTxNtEFx/nA+0dpxLTueJz+CEQFdp3haQzUXtYKqE5KZCCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8f39ccdd4c3fcf4c71ed9e9708449a07fe575cb85043f2ee90ad399e7bc72f25","last_reissued_at":"2026-07-05T06:33:07.084123Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:33:07.084123Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FACADE: A Framework for Adversarial Circuit Anomaly Detection and Evaluation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Andres Carranza, Arnuv Tandon, Dhruv Pai, Rylan Schaeffer, Sanmi Koyejo","submitted_at":"2023-07-20T04:00:37Z","abstract_excerpt":"We present FACADE, a novel probabilistic and geometric framework designed for unsupervised mechanistic anomaly detection in deep neural networks. Its primary goal is advancing the understanding and mitigation of adversarial attacks. FACADE aims to generate probabilistic distributions over circuits, which provide critical insights to their contribution to changes in the manifold properties of pseudo-classes, or high-dimensional modes in activation space, yielding a powerful tool for uncovering and combating adversarial attacks. Our approach seeks to improve model robustness, enhance scalable mo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2307.10563","kind":"arxiv","version":1},"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/2307.10563/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":"2307.10563","created_at":"2026-07-05T06:33:07.084194+00:00"},{"alias_kind":"arxiv_version","alias_value":"2307.10563v1","created_at":"2026-07-05T06:33:07.084194+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2307.10563","created_at":"2026-07-05T06:33:07.084194+00:00"},{"alias_kind":"pith_short_12","alias_value":"R444ZXKMH7HU","created_at":"2026-07-05T06:33:07.084194+00:00"},{"alias_kind":"pith_short_16","alias_value":"R444ZXKMH7HUY4PN","created_at":"2026-07-05T06:33:07.084194+00:00"},{"alias_kind":"pith_short_8","alias_value":"R444ZXKM","created_at":"2026-07-05T06:33:07.084194+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.12809","citing_title":"Correcting Influence: Unboxing LLM Outputs with Orthogonal Latent Spaces","ref_index":73,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/R444ZXKMH7HUY4PNT2LQQRE2A7","json":"https://pith.science/pith/R444ZXKMH7HUY4PNT2LQQRE2A7.json","graph_json":"https://pith.science/api/pith-number/R444ZXKMH7HUY4PNT2LQQRE2A7/graph.json","events_json":"https://pith.science/api/pith-number/R444ZXKMH7HUY4PNT2LQQRE2A7/events.json","paper":"https://pith.science/paper/R444ZXKM"},"agent_actions":{"view_html":"https://pith.science/pith/R444ZXKMH7HUY4PNT2LQQRE2A7","download_json":"https://pith.science/pith/R444ZXKMH7HUY4PNT2LQQRE2A7.json","view_paper":"https://pith.science/paper/R444ZXKM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2307.10563&json=true","fetch_graph":"https://pith.science/api/pith-number/R444ZXKMH7HUY4PNT2LQQRE2A7/graph.json","fetch_events":"https://pith.science/api/pith-number/R444ZXKMH7HUY4PNT2LQQRE2A7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/R444ZXKMH7HUY4PNT2LQQRE2A7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/R444ZXKMH7HUY4PNT2LQQRE2A7/action/storage_attestation","attest_author":"https://pith.science/pith/R444ZXKMH7HUY4PNT2LQQRE2A7/action/author_attestation","sign_citation":"https://pith.science/pith/R444ZXKMH7HUY4PNT2LQQRE2A7/action/citation_signature","submit_replication":"https://pith.science/pith/R444ZXKMH7HUY4PNT2LQQRE2A7/action/replication_record"}},"created_at":"2026-07-05T06:33:07.084194+00:00","updated_at":"2026-07-05T06:33:07.084194+00:00"}