{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:B637LZBZXUC7SHEPAH5ZQRTZFR","short_pith_number":"pith:B637LZBZ","schema_version":"1.0","canonical_sha256":"0fb7f5e439bd05f91c8f01fb9846792c5431c47f8ff69bcb7ffb7cad3716fe65","source":{"kind":"arxiv","id":"2509.22015","version":2},"attestation_state":"computed","paper":{"title":"Concept-SAE: A Controllable and Invertible Concept Interface for Sparse Autoencoders","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chenchen Zhao, Jianrong Ding, Muxi Chen, Qiang Xu","submitted_at":"2025-09-26T07:51:03Z","abstract_excerpt":"Standard Sparse Autoencoders (SAEs) excel at discovering a dictionary of a model's learned features, providing a powerful lens for passive feature discovery. However, this passive nature makes it difficult to systematically evaluate or analyze concepts that users explicitly care about. We introduce Concept-SAE, a framework that augments SAEs with a structured and controllable interface for probing user-defined concepts. Concept-SAE decomposes an activation subspace into two orthogonal components: Concept Tokens, which are aligned to externally specified semantics through dual supervision on bo"},"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":"2509.22015","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-09-26T07:51:03Z","cross_cats_sorted":[],"title_canon_sha256":"69633d46a3367be78ac1bd796c8aef6a85b048adf8ad8feba46316fed0c14269","abstract_canon_sha256":"db53a1a72951a8ebf203c173c5c92156b6ca2cf307b92ced412120aea8a2c799"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-05T01:15:16.402533Z","signature_b64":"d6G8+kj4TBko8xLFkUnNcuuIlDfwADbRjNmnNGZ7ZbHIPXFqAeLBxd32W1fAIv1u5S76KbzsbhGPuqlmd1+eAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0fb7f5e439bd05f91c8f01fb9846792c5431c47f8ff69bcb7ffb7cad3716fe65","last_reissued_at":"2026-06-05T01:15:16.401693Z","signature_status":"signed_v1","first_computed_at":"2026-06-05T01:15:16.401693Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Concept-SAE: A Controllable and Invertible Concept Interface for Sparse Autoencoders","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chenchen Zhao, Jianrong Ding, Muxi Chen, Qiang Xu","submitted_at":"2025-09-26T07:51:03Z","abstract_excerpt":"Standard Sparse Autoencoders (SAEs) excel at discovering a dictionary of a model's learned features, providing a powerful lens for passive feature discovery. However, this passive nature makes it difficult to systematically evaluate or analyze concepts that users explicitly care about. We introduce Concept-SAE, a framework that augments SAEs with a structured and controllable interface for probing user-defined concepts. Concept-SAE decomposes an activation subspace into two orthogonal components: Concept Tokens, which are aligned to externally specified semantics through dual supervision on bo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.22015","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/2509.22015/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":"2509.22015","created_at":"2026-06-05T01:15:16.401789+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.22015v2","created_at":"2026-06-05T01:15:16.401789+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.22015","created_at":"2026-06-05T01:15:16.401789+00:00"},{"alias_kind":"pith_short_12","alias_value":"B637LZBZXUC7","created_at":"2026-06-05T01:15:16.401789+00:00"},{"alias_kind":"pith_short_16","alias_value":"B637LZBZXUC7SHEP","created_at":"2026-06-05T01:15:16.401789+00:00"},{"alias_kind":"pith_short_8","alias_value":"B637LZBZ","created_at":"2026-06-05T01:15:16.401789+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/B637LZBZXUC7SHEPAH5ZQRTZFR","json":"https://pith.science/pith/B637LZBZXUC7SHEPAH5ZQRTZFR.json","graph_json":"https://pith.science/api/pith-number/B637LZBZXUC7SHEPAH5ZQRTZFR/graph.json","events_json":"https://pith.science/api/pith-number/B637LZBZXUC7SHEPAH5ZQRTZFR/events.json","paper":"https://pith.science/paper/B637LZBZ"},"agent_actions":{"view_html":"https://pith.science/pith/B637LZBZXUC7SHEPAH5ZQRTZFR","download_json":"https://pith.science/pith/B637LZBZXUC7SHEPAH5ZQRTZFR.json","view_paper":"https://pith.science/paper/B637LZBZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.22015&json=true","fetch_graph":"https://pith.science/api/pith-number/B637LZBZXUC7SHEPAH5ZQRTZFR/graph.json","fetch_events":"https://pith.science/api/pith-number/B637LZBZXUC7SHEPAH5ZQRTZFR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/B637LZBZXUC7SHEPAH5ZQRTZFR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/B637LZBZXUC7SHEPAH5ZQRTZFR/action/storage_attestation","attest_author":"https://pith.science/pith/B637LZBZXUC7SHEPAH5ZQRTZFR/action/author_attestation","sign_citation":"https://pith.science/pith/B637LZBZXUC7SHEPAH5ZQRTZFR/action/citation_signature","submit_replication":"https://pith.science/pith/B637LZBZXUC7SHEPAH5ZQRTZFR/action/replication_record"}},"created_at":"2026-06-05T01:15:16.401789+00:00","updated_at":"2026-06-05T01:15:16.401789+00:00"}