{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:KZUYCD6XCV6EPWTCRYYNN4NGGX","short_pith_number":"pith:KZUYCD6X","schema_version":"1.0","canonical_sha256":"5669810fd7157c47da628e30d6f1a635e6695123c6752ab4771c8325cf1654f8","source":{"kind":"arxiv","id":"2608.08383","version":1},"attestation_state":"computed","paper":{"title":"Safety Cost of Steering Vectors Is Separable and Reducible","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Gjergji Kasneci, Yuxiao Li","submitted_at":"2026-08-09T00:29:54Z","abstract_excerpt":"Steering vectors are a lightweight tool for controlling LLM behavior. However, emerging evidence shows that steering vectors can unintentionally compromise a model's safety mechanisms and increase compliance with harmful requests, while no effective mitigation yet exists. In this work, we show that this safety degradation arises from a separable component in the vector that disrupts the model's safety mechanisms but contributes little to the steering objective. We identify and remove this safety-degrading component, formulating the task as a constrained optimization problem solved through prim"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2608.08383","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-08-09T00:29:54Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"3cd01677d1b411fe489369fa572567775253e31fe72f46fe5317c328ab5da264","abstract_canon_sha256":"f0c786f6ae1b20b46047383b8b24e9460e2420515f1aa372e520c77c4cab5493"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-11T01:22:29.580492Z","signature_b64":"OUDNDxj0mC/nGBLIZXrTPb7EYJztZzriX87s/SSZavyoDHihOp1/x6eAVDcHJZmdPQG4/6EIP6pirnTfZuHBDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5669810fd7157c47da628e30d6f1a635e6695123c6752ab4771c8325cf1654f8","last_reissued_at":"2026-08-11T01:22:29.578054Z","signature_status":"signed_v1","first_computed_at":"2026-08-11T01:22:29.578054Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Safety Cost of Steering Vectors Is Separable and Reducible","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Gjergji Kasneci, Yuxiao Li","submitted_at":"2026-08-09T00:29:54Z","abstract_excerpt":"Steering vectors are a lightweight tool for controlling LLM behavior. However, emerging evidence shows that steering vectors can unintentionally compromise a model's safety mechanisms and increase compliance with harmful requests, while no effective mitigation yet exists. In this work, we show that this safety degradation arises from a separable component in the vector that disrupts the model's safety mechanisms but contributes little to the steering objective. We identify and remove this safety-degrading component, formulating the task as a constrained optimization problem solved through prim"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.08383","kind":"arxiv","version":1},"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/2608.08383/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2608.08383","created_at":"2026-08-11T01:22:29.578827+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.08383v1","created_at":"2026-08-11T01:22:29.578827+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.08383","created_at":"2026-08-11T01:22:29.578827+00:00"},{"alias_kind":"pith_short_12","alias_value":"KZUYCD6XCV6E","created_at":"2026-08-11T01:22:29.578827+00:00"},{"alias_kind":"pith_short_16","alias_value":"KZUYCD6XCV6EPWTC","created_at":"2026-08-11T01:22:29.578827+00:00"},{"alias_kind":"pith_short_8","alias_value":"KZUYCD6X","created_at":"2026-08-11T01:22:29.578827+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KZUYCD6XCV6EPWTCRYYNN4NGGX","json":"https://pith.science/pith/KZUYCD6XCV6EPWTCRYYNN4NGGX.json","graph_json":"https://pith.science/api/pith-number/KZUYCD6XCV6EPWTCRYYNN4NGGX/graph.json","events_json":"https://pith.science/api/pith-number/KZUYCD6XCV6EPWTCRYYNN4NGGX/events.json","paper":"https://pith.science/paper/KZUYCD6X"},"agent_actions":{"view_html":"https://pith.science/pith/KZUYCD6XCV6EPWTCRYYNN4NGGX","download_json":"https://pith.science/pith/KZUYCD6XCV6EPWTCRYYNN4NGGX.json","view_paper":"https://pith.science/paper/KZUYCD6X","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.08383&json=true","fetch_graph":"https://pith.science/api/pith-number/KZUYCD6XCV6EPWTCRYYNN4NGGX/graph.json","fetch_events":"https://pith.science/api/pith-number/KZUYCD6XCV6EPWTCRYYNN4NGGX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KZUYCD6XCV6EPWTCRYYNN4NGGX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KZUYCD6XCV6EPWTCRYYNN4NGGX/action/storage_attestation","attest_author":"https://pith.science/pith/KZUYCD6XCV6EPWTCRYYNN4NGGX/action/author_attestation","sign_citation":"https://pith.science/pith/KZUYCD6XCV6EPWTCRYYNN4NGGX/action/citation_signature","submit_replication":"https://pith.science/pith/KZUYCD6XCV6EPWTCRYYNN4NGGX/action/replication_record"}},"created_at":"2026-08-11T01:22:29.578827+00:00","updated_at":"2026-08-11T01:22:29.578827+00:00"}