{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:UDD7EK2NNZHLRQNKQTCNNUXDYM","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":"9c49b1f430dfa4ed2aec184577d3684207eaace50353c5a48f0da2a8e9d6d5f7","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-07-08T23:15:26Z","title_canon_sha256":"5d76f34bd3db5c272c9b292890c12f58fbf5654f6eb2bc8b8fa3eca1f66e29d3"},"schema_version":"1.0","source":{"id":"2207.04153","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2207.04153","created_at":"2026-07-05T06:21:48Z"},{"alias_kind":"arxiv_version","alias_value":"2207.04153v3","created_at":"2026-07-05T06:21:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2207.04153","created_at":"2026-07-05T06:21:48Z"},{"alias_kind":"pith_short_12","alias_value":"UDD7EK2NNZHL","created_at":"2026-07-05T06:21:48Z"},{"alias_kind":"pith_short_16","alias_value":"UDD7EK2NNZHLRQNK","created_at":"2026-07-05T06:21:48Z"},{"alias_kind":"pith_short_8","alias_value":"UDD7EK2N","created_at":"2026-07-05T06:21:48Z"}],"graph_snapshots":[{"event_id":"sha256:70a55d875c9ecc2a3da3acb3cd1f5f05e3f8feee22bf1745f0f7aeb7635fa795","target":"graph","created_at":"2026-07-05T06:21:48Z","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/2207.04153/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Neural network models trained on text data have been found to encode undesirable linguistic or sensitive concepts in their representation. Removing such concepts is non-trivial because of a complex relationship between the concept, text input, and the learnt representation. Recent work has proposed post-hoc and adversarial methods to remove such unwanted concepts from a model's representation. Through an extensive theoretical and empirical analysis, we show that these methods can be counter-productive: they are unable to remove the concepts entirely, and in the worst case may end up destroying","authors_text":"Abhinav Kumar, Amit Sharma, Chenhao Tan","cross_cats":["cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-07-08T23:15:26Z","title":"Probing Classifiers are Unreliable for Concept Removal and Detection"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2207.04153","kind":"arxiv","version":3},"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:d7fceef50c3e64daca60d2c3c204da991b1c30b87e0fb948c59e493c8279d8ba","target":"record","created_at":"2026-07-05T06:21:48Z","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":"9c49b1f430dfa4ed2aec184577d3684207eaace50353c5a48f0da2a8e9d6d5f7","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-07-08T23:15:26Z","title_canon_sha256":"5d76f34bd3db5c272c9b292890c12f58fbf5654f6eb2bc8b8fa3eca1f66e29d3"},"schema_version":"1.0","source":{"id":"2207.04153","kind":"arxiv","version":3}},"canonical_sha256":"a0c7f22b4d6e4eb8c1aa84c4d6d2e3c32884202ea5bec8d382ac3cd197bcfefc","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a0c7f22b4d6e4eb8c1aa84c4d6d2e3c32884202ea5bec8d382ac3cd197bcfefc","first_computed_at":"2026-07-05T06:21:48.125431Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T06:21:48.125431Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"rOViFGIl8aRY7YMb4SuNfNp2m5pitlPYrObTg9sPf1m5PQwdkFEiFkVZvdJKmy0T2rI5qTGzvxOa0sc6ov/7DQ==","signature_status":"signed_v1","signed_at":"2026-07-05T06:21:48.125841Z","signed_message":"canonical_sha256_bytes"},"source_id":"2207.04153","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:d7fceef50c3e64daca60d2c3c204da991b1c30b87e0fb948c59e493c8279d8ba","sha256:70a55d875c9ecc2a3da3acb3cd1f5f05e3f8feee22bf1745f0f7aeb7635fa795"],"state_sha256":"e45823d4b79ce6cb42e69e389dc1b42295d2d2f973c36ac8d3d585acb1e672bb"}