{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:CPYSSEOTEOUOM6F4FXPC47H6OI","short_pith_number":"pith:CPYSSEOT","canonical_record":{"source":{"id":"2411.09003","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-11-13T20:12:55Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"6407d110e8c0f67977aca5c6a4fa7cb958a04e189615764570692396caba0add","abstract_canon_sha256":"29eca4d848c1109b0b4d234d46ec1b153980fcc869ead1203fbb6ee401feb9f6"},"schema_version":"1.0"},"canonical_sha256":"13f12911d323a8e678bc2dde2e7cfe7203aa40c3d76228ac53f7bac11b6d5381","source":{"kind":"arxiv","id":"2411.09003","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2411.09003","created_at":"2026-07-05T10:06:09Z"},{"alias_kind":"arxiv_version","alias_value":"2411.09003v3","created_at":"2026-07-05T10:06:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.09003","created_at":"2026-07-05T10:06:09Z"},{"alias_kind":"pith_short_12","alias_value":"CPYSSEOTEOUO","created_at":"2026-07-05T10:06:09Z"},{"alias_kind":"pith_short_16","alias_value":"CPYSSEOTEOUOM6F4","created_at":"2026-07-05T10:06:09Z"},{"alias_kind":"pith_short_8","alias_value":"CPYSSEOT","created_at":"2026-07-05T10:06:09Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:CPYSSEOTEOUOM6F4FXPC47H6OI","target":"record","payload":{"canonical_record":{"source":{"id":"2411.09003","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-11-13T20:12:55Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"6407d110e8c0f67977aca5c6a4fa7cb958a04e189615764570692396caba0add","abstract_canon_sha256":"29eca4d848c1109b0b4d234d46ec1b153980fcc869ead1203fbb6ee401feb9f6"},"schema_version":"1.0"},"canonical_sha256":"13f12911d323a8e678bc2dde2e7cfe7203aa40c3d76228ac53f7bac11b6d5381","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:06:09.092029Z","signature_b64":"NeyvH4N+eabW4LVziH8UahMq8ZcyRFaKb/DSsO8zqJ40/c/H2T6EsrhRNdP8iCfaiybz0nhBvZqHu++YoInPBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"13f12911d323a8e678bc2dde2e7cfe7203aa40c3d76228ac53f7bac11b6d5381","last_reissued_at":"2026-07-05T10:06:09.091491Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:06:09.091491Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2411.09003","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:06:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CL/NOKma6nvs2iUKQTqMYs+A0ZsOIGLOVz87PC7CqvQ/otwuPuhV0qmnjxVdYybccT4cVYcHoxGlldw0sFHUCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T21:31:38.345648Z"},"content_sha256":"d5da631d30cdc00f25c264b7cab56afe70b105070cdf8177b2f248e8a14753c7","schema_version":"1.0","event_id":"sha256:d5da631d30cdc00f25c264b7cab56afe70b105070cdf8177b2f248e8a14753c7"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:CPYSSEOTEOUOM6F4FXPC47H6OI","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Refusal in LLMs is an Affine Function","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Adam Scherlis, Nora Belrose, Thomas Marshall","submitted_at":"2024-11-13T20:12:55Z","abstract_excerpt":"We propose affine concept editing (ACE) as an approach for steering language models' behavior by intervening directly in activations. We begin with an affine decomposition of model activation vectors and show that prior methods for steering model behavior correspond to subsets of terms of this decomposition. We then provide a derivation of ACE and use it to control refusal behavior on ten different models, including Llama 3 70B. ACE combines affine subspace projection and activation addition to reliably control the model's refusal responses across prompt types. We evaluate the results using LL"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.09003","kind":"arxiv","version":3},"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/2411.09003/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":"2026-07-05T10:06:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ygNnnG9UhsEJy7SAbj10oEcq+8hYSgDYrNeuAWf5emx+uWvie0VGT3/LTB2ozqhzRXYBTGf/xSoQmSGe2ZRWCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T21:31:38.346136Z"},"content_sha256":"5467f0cbb01588b080bb80137a0607cc19bb6251148223297a4b920d784141f7","schema_version":"1.0","event_id":"sha256:5467f0cbb01588b080bb80137a0607cc19bb6251148223297a4b920d784141f7"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/CPYSSEOTEOUOM6F4FXPC47H6OI/bundle.json","state_url":"https://pith.science/pith/CPYSSEOTEOUOM6F4FXPC47H6OI/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/CPYSSEOTEOUOM6F4FXPC47H6OI/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":"2026-08-06T21:31:38Z","links":{"resolver":"https://pith.science/pith/CPYSSEOTEOUOM6F4FXPC47H6OI","bundle":"https://pith.science/pith/CPYSSEOTEOUOM6F4FXPC47H6OI/bundle.json","state":"https://pith.science/pith/CPYSSEOTEOUOM6F4FXPC47H6OI/state.json","well_known_bundle":"https://pith.science/.well-known/pith/CPYSSEOTEOUOM6F4FXPC47H6OI/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:CPYSSEOTEOUOM6F4FXPC47H6OI","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":"29eca4d848c1109b0b4d234d46ec1b153980fcc869ead1203fbb6ee401feb9f6","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-11-13T20:12:55Z","title_canon_sha256":"6407d110e8c0f67977aca5c6a4fa7cb958a04e189615764570692396caba0add"},"schema_version":"1.0","source":{"id":"2411.09003","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2411.09003","created_at":"2026-07-05T10:06:09Z"},{"alias_kind":"arxiv_version","alias_value":"2411.09003v3","created_at":"2026-07-05T10:06:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.09003","created_at":"2026-07-05T10:06:09Z"},{"alias_kind":"pith_short_12","alias_value":"CPYSSEOTEOUO","created_at":"2026-07-05T10:06:09Z"},{"alias_kind":"pith_short_16","alias_value":"CPYSSEOTEOUOM6F4","created_at":"2026-07-05T10:06:09Z"},{"alias_kind":"pith_short_8","alias_value":"CPYSSEOT","created_at":"2026-07-05T10:06:09Z"}],"graph_snapshots":[{"event_id":"sha256:5467f0cbb01588b080bb80137a0607cc19bb6251148223297a4b920d784141f7","target":"graph","created_at":"2026-07-05T10:06:09Z","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/2411.09003/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We propose affine concept editing (ACE) as an approach for steering language models' behavior by intervening directly in activations. We begin with an affine decomposition of model activation vectors and show that prior methods for steering model behavior correspond to subsets of terms of this decomposition. We then provide a derivation of ACE and use it to control refusal behavior on ten different models, including Llama 3 70B. ACE combines affine subspace projection and activation addition to reliably control the model's refusal responses across prompt types. We evaluate the results using LL","authors_text":"Adam Scherlis, Nora Belrose, Thomas Marshall","cross_cats":["cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-11-13T20:12:55Z","title":"Refusal in LLMs is an Affine Function"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.09003","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:d5da631d30cdc00f25c264b7cab56afe70b105070cdf8177b2f248e8a14753c7","target":"record","created_at":"2026-07-05T10:06:09Z","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":"29eca4d848c1109b0b4d234d46ec1b153980fcc869ead1203fbb6ee401feb9f6","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-11-13T20:12:55Z","title_canon_sha256":"6407d110e8c0f67977aca5c6a4fa7cb958a04e189615764570692396caba0add"},"schema_version":"1.0","source":{"id":"2411.09003","kind":"arxiv","version":3}},"canonical_sha256":"13f12911d323a8e678bc2dde2e7cfe7203aa40c3d76228ac53f7bac11b6d5381","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"13f12911d323a8e678bc2dde2e7cfe7203aa40c3d76228ac53f7bac11b6d5381","first_computed_at":"2026-07-05T10:06:09.091491Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:06:09.091491Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"NeyvH4N+eabW4LVziH8UahMq8ZcyRFaKb/DSsO8zqJ40/c/H2T6EsrhRNdP8iCfaiybz0nhBvZqHu++YoInPBA==","signature_status":"signed_v1","signed_at":"2026-07-05T10:06:09.092029Z","signed_message":"canonical_sha256_bytes"},"source_id":"2411.09003","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:d5da631d30cdc00f25c264b7cab56afe70b105070cdf8177b2f248e8a14753c7","sha256:5467f0cbb01588b080bb80137a0607cc19bb6251148223297a4b920d784141f7"],"state_sha256":"bf23caacfac0f005774eced41b86ffacc4ed1cc99aa566734e04c7d0c97135f4"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/Lw1w/hEtqurRXS0lypf+D3TS1G3KLgSxPrS68JFHkQB5VI22M/4/SJgOP5xKaY+Y6sXV9Y588zaHaojXUMzCw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-06T21:31:38.349808Z","bundle_sha256":"7546efad6c3b5415e373add2884974cec05fdde9d6abecb9ab00451067547b98"}}