{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:JOO47KF2SMPHAQGDUXVQI75L2N","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":"58fc0602bb2256abc076b10c597f94b7a74428a5c587ee41511ed9443c1df50d","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-11-16T19:32:43Z","title_canon_sha256":"bfd91cbb0da5c34427f4bfa9858718fdebba6c82bb6c39b611fba4763248748f"},"schema_version":"1.0","source":{"id":"2211.09169","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2211.09169","created_at":"2026-07-05T05:16:53Z"},{"alias_kind":"arxiv_version","alias_value":"2211.09169v1","created_at":"2026-07-05T05:16:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2211.09169","created_at":"2026-07-05T05:16:53Z"},{"alias_kind":"pith_short_12","alias_value":"JOO47KF2SMPH","created_at":"2026-07-05T05:16:53Z"},{"alias_kind":"pith_short_16","alias_value":"JOO47KF2SMPHAQGD","created_at":"2026-07-05T05:16:53Z"},{"alias_kind":"pith_short_8","alias_value":"JOO47KF2","created_at":"2026-07-05T05:16:53Z"}],"graph_snapshots":[{"event_id":"sha256:7447a00cdcf441149afdc434c41a23ffbcf93278d634ecd9fd1cef5d24e7ea48","target":"graph","created_at":"2026-07-05T05:16:53Z","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/2211.09169/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"In some neural networks, individual neurons correspond to natural ``features'' in the input. Such \\emph{monosemantic} neurons are of great help in interpretability studies, as they can be cleanly understood. In this work we report preliminary attempts to engineer monosemanticity in toy models. We find that models can be made more monosemantic without increasing the loss by just changing which local minimum the training process finds. More monosemantic loss minima have moderate negative biases, and we are able to use this fact to engineer highly monosemantic models. We are able to mechanistical","authors_text":"Adam S. Jermyn, Evan Hubinger, Nicholas Schiefer","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-11-16T19:32:43Z","title":"Engineering Monosemanticity in Toy Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2211.09169","kind":"arxiv","version":1},"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:df4d35cab3fc8b7a2023e6cf2caa0b8337ea1eeea298c7bc93e3ca8c7d464a4a","target":"record","created_at":"2026-07-05T05:16:53Z","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":"58fc0602bb2256abc076b10c597f94b7a74428a5c587ee41511ed9443c1df50d","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-11-16T19:32:43Z","title_canon_sha256":"bfd91cbb0da5c34427f4bfa9858718fdebba6c82bb6c39b611fba4763248748f"},"schema_version":"1.0","source":{"id":"2211.09169","kind":"arxiv","version":1}},"canonical_sha256":"4b9dcfa8ba931e7040c3a5eb047fabd341c50907444055986dc9ce6c90e33eda","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4b9dcfa8ba931e7040c3a5eb047fabd341c50907444055986dc9ce6c90e33eda","first_computed_at":"2026-07-05T05:16:53.409409Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T05:16:53.409409Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"qG+5loiq2MM8oyD3qc8jDXBA4ZHIDGUTboc9UcInfIXn3txE7R79vlCfhQiVZVsIirqCoxFWTdV0qYjRfNCPAw==","signature_status":"signed_v1","signed_at":"2026-07-05T05:16:53.409805Z","signed_message":"canonical_sha256_bytes"},"source_id":"2211.09169","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:df4d35cab3fc8b7a2023e6cf2caa0b8337ea1eeea298c7bc93e3ca8c7d464a4a","sha256:7447a00cdcf441149afdc434c41a23ffbcf93278d634ecd9fd1cef5d24e7ea48"],"state_sha256":"4ca96a0ef16cbe7c4f7a355591dc8f5bcb995fabbf3a15f6036c6e06f02d5555"}