{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:4JLGDGR57SAGKPW4JFARF6SUWH","short_pith_number":"pith:4JLGDGR5","schema_version":"1.0","canonical_sha256":"e256619a3dfc80653edc494112fa54b1ecb40687fddaf34f9868f709dacacf59","source":{"kind":"arxiv","id":"2407.13594","version":2},"attestation_state":"computed","paper":{"title":"Validating Mechanistic Interpretations: An Axiomatic Approach","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Corina S. Pasareanu, Nils Palumbo, Ravi Mangal, Saranya Vijayakumar, Somesh Jha, Zifan Wang","submitted_at":"2024-07-18T15:32:44Z","abstract_excerpt":"Mechanistic interpretability aims to reverse engineer the computation performed by a neural network in terms of its internal components. Although there is a growing body of research on mechanistic interpretation of neural networks, the notion of a mechanistic interpretation itself is often ad-hoc. Inspired by the notion of abstract interpretation from the program analysis literature that aims to develop approximate semantics for programs, we give a set of axioms that formally characterize a mechanistic interpretation as a description that approximately captures the semantics of the neural netw"},"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":"2407.13594","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-07-18T15:32:44Z","cross_cats_sorted":[],"title_canon_sha256":"f2910a459d14a0f0f07ca5582e6de382559b2dc5cd13034ea4bafded70e1bb12","abstract_canon_sha256":"e369a1345d14cd9a2edee98687bdf303462371de2eb2ccb330e5ce222d117c2c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:25:02.919406Z","signature_b64":"L9OQ/9tEZ97WQpgKOO7EDFNX1ENRF9kvnxXJJ9fCzG7egmZGmVoEjW4I+hjfjT8mxjBIVk+Kraf+LO9uiQ/hBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e256619a3dfc80653edc494112fa54b1ecb40687fddaf34f9868f709dacacf59","last_reissued_at":"2026-07-05T11:25:02.918899Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:25:02.918899Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Validating Mechanistic Interpretations: An Axiomatic Approach","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Corina S. Pasareanu, Nils Palumbo, Ravi Mangal, Saranya Vijayakumar, Somesh Jha, Zifan Wang","submitted_at":"2024-07-18T15:32:44Z","abstract_excerpt":"Mechanistic interpretability aims to reverse engineer the computation performed by a neural network in terms of its internal components. Although there is a growing body of research on mechanistic interpretation of neural networks, the notion of a mechanistic interpretation itself is often ad-hoc. Inspired by the notion of abstract interpretation from the program analysis literature that aims to develop approximate semantics for programs, we give a set of axioms that formally characterize a mechanistic interpretation as a description that approximately captures the semantics of the neural netw"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.13594","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/2407.13594/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":"2407.13594","created_at":"2026-07-05T11:25:02.918961+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.13594v2","created_at":"2026-07-05T11:25:02.918961+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.13594","created_at":"2026-07-05T11:25:02.918961+00:00"},{"alias_kind":"pith_short_12","alias_value":"4JLGDGR57SAG","created_at":"2026-07-05T11:25:02.918961+00:00"},{"alias_kind":"pith_short_16","alias_value":"4JLGDGR57SAGKPW4","created_at":"2026-07-05T11:25:02.918961+00:00"},{"alias_kind":"pith_short_8","alias_value":"4JLGDGR5","created_at":"2026-07-05T11:25:02.918961+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.17704","citing_title":"Toy Combinatorial Interpretability Models Reveal Lottery Tickets in Early Feature Space","ref_index":60,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4JLGDGR57SAGKPW4JFARF6SUWH","json":"https://pith.science/pith/4JLGDGR57SAGKPW4JFARF6SUWH.json","graph_json":"https://pith.science/api/pith-number/4JLGDGR57SAGKPW4JFARF6SUWH/graph.json","events_json":"https://pith.science/api/pith-number/4JLGDGR57SAGKPW4JFARF6SUWH/events.json","paper":"https://pith.science/paper/4JLGDGR5"},"agent_actions":{"view_html":"https://pith.science/pith/4JLGDGR57SAGKPW4JFARF6SUWH","download_json":"https://pith.science/pith/4JLGDGR57SAGKPW4JFARF6SUWH.json","view_paper":"https://pith.science/paper/4JLGDGR5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.13594&json=true","fetch_graph":"https://pith.science/api/pith-number/4JLGDGR57SAGKPW4JFARF6SUWH/graph.json","fetch_events":"https://pith.science/api/pith-number/4JLGDGR57SAGKPW4JFARF6SUWH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4JLGDGR57SAGKPW4JFARF6SUWH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4JLGDGR57SAGKPW4JFARF6SUWH/action/storage_attestation","attest_author":"https://pith.science/pith/4JLGDGR57SAGKPW4JFARF6SUWH/action/author_attestation","sign_citation":"https://pith.science/pith/4JLGDGR57SAGKPW4JFARF6SUWH/action/citation_signature","submit_replication":"https://pith.science/pith/4JLGDGR57SAGKPW4JFARF6SUWH/action/replication_record"}},"created_at":"2026-07-05T11:25:02.918961+00:00","updated_at":"2026-07-05T11:25:02.918961+00:00"}