{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:B3AX7MYI5366C6MXQQYIFBDOZN","short_pith_number":"pith:B3AX7MYI","schema_version":"1.0","canonical_sha256":"0ec17fb308eefde17997843082846ecb41c1c6c946bdb7b023b99341947350af","source":{"kind":"arxiv","id":"2602.07356","version":2},"attestation_state":"computed","paper":{"title":"Controllable Value Alignment in Large Language Models through Neuron-Level Editing","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Fengbin Zhu, Jilong Liu, Junwei Li, Le Wu, Richang Hong, Tat-Seng Chua, Weibiao Huang, Yihui Wang, Yonghui Yang","submitted_at":"2026-02-07T04:35:16Z","abstract_excerpt":"Aligning large language models (LLMs) with human values has become increasingly important as their influence on human behavior and decision-making expands. However, existing steering-based alignment methods suffer from limited controllability: steering a target value often unintentionally activates other, non-target values. To characterize this limitation, we introduce value leakage, a diagnostic notion that captures the unintended activation of non-target values during value steering, along with a normalized leakage metric grounded in Schwartz's value theory. In light of this analysis, we pro"},"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":"2602.07356","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-02-07T04:35:16Z","cross_cats_sorted":[],"title_canon_sha256":"6dc7054d6f118f4aacd5c617d28b5a56e149d4bd8b7013bb30696680bae3d6f0","abstract_canon_sha256":"43a669b59dc8ab421218f6db2b4ce2f3a79bdd36b7e90b661c1b27b77e56a855"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T02:04:51.115460Z","signature_b64":"MZtW+h8ndH+d6lHyyPbMgJ3LhTwdj6+p5eTGEtfSkojg1P3IUq63RPPu2ib729zCj9Xx3BRKlzlwyyHlKuBdAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0ec17fb308eefde17997843082846ecb41c1c6c946bdb7b023b99341947350af","last_reissued_at":"2026-06-02T02:04:51.114940Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T02:04:51.114940Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Controllable Value Alignment in Large Language Models through Neuron-Level Editing","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Fengbin Zhu, Jilong Liu, Junwei Li, Le Wu, Richang Hong, Tat-Seng Chua, Weibiao Huang, Yihui Wang, Yonghui Yang","submitted_at":"2026-02-07T04:35:16Z","abstract_excerpt":"Aligning large language models (LLMs) with human values has become increasingly important as their influence on human behavior and decision-making expands. However, existing steering-based alignment methods suffer from limited controllability: steering a target value often unintentionally activates other, non-target values. To characterize this limitation, we introduce value leakage, a diagnostic notion that captures the unintended activation of non-target values during value steering, along with a normalized leakage metric grounded in Schwartz's value theory. In light of this analysis, we pro"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.07356","kind":"arxiv","version":2},"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/2602.07356/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":"2602.07356","created_at":"2026-06-02T02:04:51.115003+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.07356v2","created_at":"2026-06-02T02:04:51.115003+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.07356","created_at":"2026-06-02T02:04:51.115003+00:00"},{"alias_kind":"pith_short_12","alias_value":"B3AX7MYI5366","created_at":"2026-06-02T02:04:51.115003+00:00"},{"alias_kind":"pith_short_16","alias_value":"B3AX7MYI5366C6MX","created_at":"2026-06-02T02:04:51.115003+00:00"},{"alias_kind":"pith_short_8","alias_value":"B3AX7MYI","created_at":"2026-06-02T02:04:51.115003+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/B3AX7MYI5366C6MXQQYIFBDOZN","json":"https://pith.science/pith/B3AX7MYI5366C6MXQQYIFBDOZN.json","graph_json":"https://pith.science/api/pith-number/B3AX7MYI5366C6MXQQYIFBDOZN/graph.json","events_json":"https://pith.science/api/pith-number/B3AX7MYI5366C6MXQQYIFBDOZN/events.json","paper":"https://pith.science/paper/B3AX7MYI"},"agent_actions":{"view_html":"https://pith.science/pith/B3AX7MYI5366C6MXQQYIFBDOZN","download_json":"https://pith.science/pith/B3AX7MYI5366C6MXQQYIFBDOZN.json","view_paper":"https://pith.science/paper/B3AX7MYI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.07356&json=true","fetch_graph":"https://pith.science/api/pith-number/B3AX7MYI5366C6MXQQYIFBDOZN/graph.json","fetch_events":"https://pith.science/api/pith-number/B3AX7MYI5366C6MXQQYIFBDOZN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/B3AX7MYI5366C6MXQQYIFBDOZN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/B3AX7MYI5366C6MXQQYIFBDOZN/action/storage_attestation","attest_author":"https://pith.science/pith/B3AX7MYI5366C6MXQQYIFBDOZN/action/author_attestation","sign_citation":"https://pith.science/pith/B3AX7MYI5366C6MXQQYIFBDOZN/action/citation_signature","submit_replication":"https://pith.science/pith/B3AX7MYI5366C6MXQQYIFBDOZN/action/replication_record"}},"created_at":"2026-06-02T02:04:51.115003+00:00","updated_at":"2026-06-02T02:04:51.115003+00:00"}