{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:GBWTCE3MGKAVOM3KZV6CLBJKOR","short_pith_number":"pith:GBWTCE3M","canonical_record":{"source":{"id":"2201.12091","kind":"arxiv","version":8},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-01-28T13:00:17Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"4c4f9624cc6bf448d5000f1d353cd104d75c059de95372899c8243423aa4a4a8","abstract_canon_sha256":"4fa19b63f3ddf2c4caf9d378ce23cbc456084ea4a5c1b451565cc9135d3bf686"},"schema_version":"1.0"},"canonical_sha256":"306d31136c328157336acd7c25852a745e4a94cceeefc1f84bc4710a242db2cb","source":{"kind":"arxiv","id":"2201.12091","version":8},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2201.12091","created_at":"2026-07-05T09:49:57Z"},{"alias_kind":"arxiv_version","alias_value":"2201.12091v8","created_at":"2026-07-05T09:49:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2201.12091","created_at":"2026-07-05T09:49:57Z"},{"alias_kind":"pith_short_12","alias_value":"GBWTCE3MGKAV","created_at":"2026-07-05T09:49:57Z"},{"alias_kind":"pith_short_16","alias_value":"GBWTCE3MGKAVOM3K","created_at":"2026-07-05T09:49:57Z"},{"alias_kind":"pith_short_8","alias_value":"GBWTCE3M","created_at":"2026-07-05T09:49:57Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:GBWTCE3MGKAVOM3KZV6CLBJKOR","target":"record","payload":{"canonical_record":{"source":{"id":"2201.12091","kind":"arxiv","version":8},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-01-28T13:00:17Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"4c4f9624cc6bf448d5000f1d353cd104d75c059de95372899c8243423aa4a4a8","abstract_canon_sha256":"4fa19b63f3ddf2c4caf9d378ce23cbc456084ea4a5c1b451565cc9135d3bf686"},"schema_version":"1.0"},"canonical_sha256":"306d31136c328157336acd7c25852a745e4a94cceeefc1f84bc4710a242db2cb","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:49:57.328155Z","signature_b64":"Q3pf6qMtqN43/z06PKRoYl1pu5HdADmuU/tc6DUejEX9orr94bGpiTRSQGN4nrzaF0iXJree9uLXHmkIQojtDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"306d31136c328157336acd7c25852a745e4a94cceeefc1f84bc4710a242db2cb","last_reissued_at":"2026-07-05T09:49:57.327604Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:49:57.327604Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2201.12091","source_version":8,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:49:57Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"SEbciUXqfZfYK0IEHQNMD8vXV1g2gWl8LSOLo7XzNovoN/jne5qJ+ki0prCX6oXkQr3JAD3phCYsgtjAqVJnCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-11T04:50:26.647246Z"},"content_sha256":"1530f5cdefb812e166b71559d9c0679dab862cf75bdd47b2143e5345db2e44ac","schema_version":"1.0","event_id":"sha256:1530f5cdefb812e166b71559d9c0679dab862cf75bdd47b2143e5345db2e44ac"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:GBWTCE3MGKAVOM3KZV6CLBJKOR","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Linear Adversarial Concept Erasure","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Michael Twiton, Ryan Cotterell, Shauli Ravfogel, Yoav Goldberg","submitted_at":"2022-01-28T13:00:17Z","abstract_excerpt":"Modern neural models trained on textual data rely on pre-trained representations that emerge without direct supervision. As these representations are increasingly being used in real-world applications, the inability to \\emph{control} their content becomes an increasingly important problem. We formulate the problem of identifying and erasing a linear subspace that corresponds to a given concept, in order to prevent linear predictors from recovering the concept. We model this problem as a constrained, linear maximin game, and show that existing solutions are generally not optimal for this task. "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2201.12091","kind":"arxiv","version":8},"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/2201.12091/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:49:57Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"knFwYIEUdn67k4E+L0+ziXCygKo6jykDKy26MDVcapwT7FujaZBgUEtjk+OZcgRNCvROuXCDSaIZj4dRJt8TAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-11T04:50:26.647795Z"},"content_sha256":"3ed30237d9754fedb08f9df9a27e4ce089918e879baf124a77f5e6cd01a9b696","schema_version":"1.0","event_id":"sha256:3ed30237d9754fedb08f9df9a27e4ce089918e879baf124a77f5e6cd01a9b696"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/GBWTCE3MGKAVOM3KZV6CLBJKOR/bundle.json","state_url":"https://pith.science/pith/GBWTCE3MGKAVOM3KZV6CLBJKOR/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/GBWTCE3MGKAVOM3KZV6CLBJKOR/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-11T04:50:26Z","links":{"resolver":"https://pith.science/pith/GBWTCE3MGKAVOM3KZV6CLBJKOR","bundle":"https://pith.science/pith/GBWTCE3MGKAVOM3KZV6CLBJKOR/bundle.json","state":"https://pith.science/pith/GBWTCE3MGKAVOM3KZV6CLBJKOR/state.json","well_known_bundle":"https://pith.science/.well-known/pith/GBWTCE3MGKAVOM3KZV6CLBJKOR/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:GBWTCE3MGKAVOM3KZV6CLBJKOR","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"4fa19b63f3ddf2c4caf9d378ce23cbc456084ea4a5c1b451565cc9135d3bf686","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-01-28T13:00:17Z","title_canon_sha256":"4c4f9624cc6bf448d5000f1d353cd104d75c059de95372899c8243423aa4a4a8"},"schema_version":"1.0","source":{"id":"2201.12091","kind":"arxiv","version":8}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2201.12091","created_at":"2026-07-05T09:49:57Z"},{"alias_kind":"arxiv_version","alias_value":"2201.12091v8","created_at":"2026-07-05T09:49:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2201.12091","created_at":"2026-07-05T09:49:57Z"},{"alias_kind":"pith_short_12","alias_value":"GBWTCE3MGKAV","created_at":"2026-07-05T09:49:57Z"},{"alias_kind":"pith_short_16","alias_value":"GBWTCE3MGKAVOM3K","created_at":"2026-07-05T09:49:57Z"},{"alias_kind":"pith_short_8","alias_value":"GBWTCE3M","created_at":"2026-07-05T09:49:57Z"}],"graph_snapshots":[{"event_id":"sha256:3ed30237d9754fedb08f9df9a27e4ce089918e879baf124a77f5e6cd01a9b696","target":"graph","created_at":"2026-07-05T09:49:57Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2201.12091/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Modern neural models trained on textual data rely on pre-trained representations that emerge without direct supervision. As these representations are increasingly being used in real-world applications, the inability to \\emph{control} their content becomes an increasingly important problem. We formulate the problem of identifying and erasing a linear subspace that corresponds to a given concept, in order to prevent linear predictors from recovering the concept. We model this problem as a constrained, linear maximin game, and show that existing solutions are generally not optimal for this task. ","authors_text":"Michael Twiton, Ryan Cotterell, Shauli Ravfogel, Yoav Goldberg","cross_cats":["cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-01-28T13:00:17Z","title":"Linear Adversarial Concept Erasure"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2201.12091","kind":"arxiv","version":8},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:1530f5cdefb812e166b71559d9c0679dab862cf75bdd47b2143e5345db2e44ac","target":"record","created_at":"2026-07-05T09:49:57Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"4fa19b63f3ddf2c4caf9d378ce23cbc456084ea4a5c1b451565cc9135d3bf686","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-01-28T13:00:17Z","title_canon_sha256":"4c4f9624cc6bf448d5000f1d353cd104d75c059de95372899c8243423aa4a4a8"},"schema_version":"1.0","source":{"id":"2201.12091","kind":"arxiv","version":8}},"canonical_sha256":"306d31136c328157336acd7c25852a745e4a94cceeefc1f84bc4710a242db2cb","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"306d31136c328157336acd7c25852a745e4a94cceeefc1f84bc4710a242db2cb","first_computed_at":"2026-07-05T09:49:57.327604Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:49:57.327604Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Q3pf6qMtqN43/z06PKRoYl1pu5HdADmuU/tc6DUejEX9orr94bGpiTRSQGN4nrzaF0iXJree9uLXHmkIQojtDQ==","signature_status":"signed_v1","signed_at":"2026-07-05T09:49:57.328155Z","signed_message":"canonical_sha256_bytes"},"source_id":"2201.12091","source_kind":"arxiv","source_version":8}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1530f5cdefb812e166b71559d9c0679dab862cf75bdd47b2143e5345db2e44ac","sha256:3ed30237d9754fedb08f9df9a27e4ce089918e879baf124a77f5e6cd01a9b696"],"state_sha256":"e45d94eef36081b56826a711cb7a3fe382610910c9b9f5fe9bf861aaadf7a038"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"KtSbj0iFSYL24DIzuaslEcGzlFfaxcs98k2EvZhJLuy1btchgwOqENRQSVfdnblmUctGv842uFwDKqjvwTHaAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-11T04:50:26.651838Z","bundle_sha256":"31c9b18dd5f827a38966336dd0fa0b78c586a57466b5cebf7fe8c6c8185d8d7d"}}