{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:EVSXM3KOURUVMNJ6TJTJS7YLE3","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":"ef0286ea73b187323efb6999937011732c84a0c8169441902d55817a213dcb99","cross_cats_sorted":["cs.CL","cs.CY"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-15T00:20:30Z","title_canon_sha256":"9ab3aa4670b3db5bdbbc606e8deb363af2e787964bec9975da00dbe6d909f8d7"},"schema_version":"1.0","source":{"id":"2402.09631","kind":"arxiv","version":7}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.09631","created_at":"2026-07-05T11:15:36Z"},{"alias_kind":"arxiv_version","alias_value":"2402.09631v7","created_at":"2026-07-05T11:15:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.09631","created_at":"2026-07-05T11:15:36Z"},{"alias_kind":"pith_short_12","alias_value":"EVSXM3KOURUV","created_at":"2026-07-05T11:15:36Z"},{"alias_kind":"pith_short_16","alias_value":"EVSXM3KOURUVMNJ6","created_at":"2026-07-05T11:15:36Z"},{"alias_kind":"pith_short_8","alias_value":"EVSXM3KO","created_at":"2026-07-05T11:15:36Z"}],"graph_snapshots":[{"event_id":"sha256:2ea3dc61ce4493f3af82c6d5c6e3f8ed387ab7b976abd1a5fcef44532c517aed","target":"graph","created_at":"2026-07-05T11:15:36Z","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/2402.09631/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Language models often exhibit undesirable behavior, e.g., generating toxic or gender-biased text. In the case of neural language models, an encoding of the undesirable behavior is often present in the model's representations. Thus, one natural (and common) approach to prevent the model from exhibiting undesirable behavior is to steer the model's representations in a manner that reduces the probability of it generating undesirable text. This paper investigates the formal and empirical properties of steering functions, i.e., transformation of the neural language model's representations that alte","authors_text":"Jonathan Herzig, Ponnurangam Kumaraguru, Roee Aharoni, Ryan Cotterell, Shashwat Singh, Shauli Ravfogel","cross_cats":["cs.CL","cs.CY"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-15T00:20:30Z","title":"Representation Surgery: Theory and Practice of Affine Steering"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.09631","kind":"arxiv","version":7},"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:bd38276644b6fa0c450236719a7d01aa3cfda13bb79cf9295c48268e8e782ff5","target":"record","created_at":"2026-07-05T11:15:36Z","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":"ef0286ea73b187323efb6999937011732c84a0c8169441902d55817a213dcb99","cross_cats_sorted":["cs.CL","cs.CY"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-15T00:20:30Z","title_canon_sha256":"9ab3aa4670b3db5bdbbc606e8deb363af2e787964bec9975da00dbe6d909f8d7"},"schema_version":"1.0","source":{"id":"2402.09631","kind":"arxiv","version":7}},"canonical_sha256":"2565766d4ea46956353e9a66997f0b26dd9a88e60bd4f1d66c3c6b2ed31ff65c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2565766d4ea46956353e9a66997f0b26dd9a88e60bd4f1d66c3c6b2ed31ff65c","first_computed_at":"2026-07-05T11:15:36.913896Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:15:36.913896Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"kJxTK2CTpZBTQuoo7Gso3o8FK0k64n8Qe02xM/Za68GslfhwIAXyyQK+MI3oz+gDyA0s0aXBXoIv6RvG8ou9CQ==","signature_status":"signed_v1","signed_at":"2026-07-05T11:15:36.914461Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.09631","source_kind":"arxiv","source_version":7}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:bd38276644b6fa0c450236719a7d01aa3cfda13bb79cf9295c48268e8e782ff5","sha256:2ea3dc61ce4493f3af82c6d5c6e3f8ed387ab7b976abd1a5fcef44532c517aed"],"state_sha256":"a5c0b8b4e3024e57b77a7b7f221ec16095aa67cd4c265c532538b939c1a74025"}