{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:FY3WBTXTO3KCLRTSIJ36OPFLRR","short_pith_number":"pith:FY3WBTXT","schema_version":"1.0","canonical_sha256":"2e3760cef376d425c6724277e73cab8c5ffebe04b5a9d91f17df0d7b9b01b76d","source":{"kind":"arxiv","id":"2607.19806","version":1},"attestation_state":"computed","paper":{"title":"OPIUM: Mitigating Steering Externalities and Over-Refusal via Dual Objective Latent Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Aadi Prasad, Arihant Rastogi, Kavin Aravindan, Krishak Aneja, Ponnurangam Kumaraguru, Saiyam Jain, Vaishnavi Shivkumar","submitted_at":"2026-07-22T06:39:03Z","abstract_excerpt":"Activation steering provides a lightweight mechanism for controlling large language models at inference time, but steering vectors can have unintended externalities: utility vectors may weaken safety behavior, while refusal vectors may induce over-refusal on benign prompts. We introduce OPIUM (Optimizing Protected Injections via Utility Manifolds), a training-free method for sanitizing steering vectors through representation matching. Given reference behaviors on two prompt sets, OPIUM optimizes a new steering vector that preserves the downstream representations induced by the desired interven"},"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":"2607.19806","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-22T06:39:03Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"60ef65731d7c61ab4033d8a9f873f7f2f0d9539aa57cb91b4f357ea2f5a183a8","abstract_canon_sha256":"7db89a8e00b93247ffd4c94dbdc1b6377ce0a4201ee73b5433340b5a31c981fa"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-23T01:24:00.294114Z","signature_b64":"2nNT5GSwjb9/FhcoBaNjdelBAyZYROQCNiLsD44G6zAtmiHD7hgdynQ+zN5cHF7zMd7x77Bv/5AVesUZmFhHAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2e3760cef376d425c6724277e73cab8c5ffebe04b5a9d91f17df0d7b9b01b76d","last_reissued_at":"2026-07-23T01:24:00.293197Z","signature_status":"signed_v1","first_computed_at":"2026-07-23T01:24:00.293197Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"OPIUM: Mitigating Steering Externalities and Over-Refusal via Dual Objective Latent Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Aadi Prasad, Arihant Rastogi, Kavin Aravindan, Krishak Aneja, Ponnurangam Kumaraguru, Saiyam Jain, Vaishnavi Shivkumar","submitted_at":"2026-07-22T06:39:03Z","abstract_excerpt":"Activation steering provides a lightweight mechanism for controlling large language models at inference time, but steering vectors can have unintended externalities: utility vectors may weaken safety behavior, while refusal vectors may induce over-refusal on benign prompts. We introduce OPIUM (Optimizing Protected Injections via Utility Manifolds), a training-free method for sanitizing steering vectors through representation matching. Given reference behaviors on two prompt sets, OPIUM optimizes a new steering vector that preserves the downstream representations induced by the desired interven"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.19806","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/2607.19806/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":"2607.19806","created_at":"2026-07-23T01:24:00.293681+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.19806v1","created_at":"2026-07-23T01:24:00.293681+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.19806","created_at":"2026-07-23T01:24:00.293681+00:00"},{"alias_kind":"pith_short_12","alias_value":"FY3WBTXTO3KC","created_at":"2026-07-23T01:24:00.293681+00:00"},{"alias_kind":"pith_short_16","alias_value":"FY3WBTXTO3KCLRTS","created_at":"2026-07-23T01:24:00.293681+00:00"},{"alias_kind":"pith_short_8","alias_value":"FY3WBTXT","created_at":"2026-07-23T01:24:00.293681+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/FY3WBTXTO3KCLRTSIJ36OPFLRR","json":"https://pith.science/pith/FY3WBTXTO3KCLRTSIJ36OPFLRR.json","graph_json":"https://pith.science/api/pith-number/FY3WBTXTO3KCLRTSIJ36OPFLRR/graph.json","events_json":"https://pith.science/api/pith-number/FY3WBTXTO3KCLRTSIJ36OPFLRR/events.json","paper":"https://pith.science/paper/FY3WBTXT"},"agent_actions":{"view_html":"https://pith.science/pith/FY3WBTXTO3KCLRTSIJ36OPFLRR","download_json":"https://pith.science/pith/FY3WBTXTO3KCLRTSIJ36OPFLRR.json","view_paper":"https://pith.science/paper/FY3WBTXT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.19806&json=true","fetch_graph":"https://pith.science/api/pith-number/FY3WBTXTO3KCLRTSIJ36OPFLRR/graph.json","fetch_events":"https://pith.science/api/pith-number/FY3WBTXTO3KCLRTSIJ36OPFLRR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FY3WBTXTO3KCLRTSIJ36OPFLRR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FY3WBTXTO3KCLRTSIJ36OPFLRR/action/storage_attestation","attest_author":"https://pith.science/pith/FY3WBTXTO3KCLRTSIJ36OPFLRR/action/author_attestation","sign_citation":"https://pith.science/pith/FY3WBTXTO3KCLRTSIJ36OPFLRR/action/citation_signature","submit_replication":"https://pith.science/pith/FY3WBTXTO3KCLRTSIJ36OPFLRR/action/replication_record"}},"created_at":"2026-07-23T01:24:00.293681+00:00","updated_at":"2026-07-23T01:24:00.293681+00:00"}