{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:WC367CFYMUU3UZ5QTA5QNTWILO","short_pith_number":"pith:WC367CFY","schema_version":"1.0","canonical_sha256":"b0b7ef88b86529ba67b0983b06cec85b9598cd77555fb69e8cea02d516429375","source":{"kind":"arxiv","id":"2410.03772","version":1},"attestation_state":"computed","paper":{"title":"Precision Knowledge Editing: Enhancing Safety in Large Language Models","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Victor Bian, Xuying Li, Yasuhiro Yoshida, Yuji Kosuga, Zhuo Li","submitted_at":"2024-10-02T23:15:53Z","abstract_excerpt":"Large language models (LLMs) have demonstrated remarkable capabilities, but they also pose risks related to the generation of toxic or harmful content. This work introduces Precision Knowledge Editing (PKE), an advanced technique that builds upon existing knowledge editing methods to more effectively identify and modify toxic parameter regions within LLMs. By leveraging neuron weight tracking and activation pathway tracing, PKE achieves finer granularity in toxic content management compared to previous methods like Detoxifying Instance Neuron Modification (DINM). Our experiments demonstrate th"},"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":"2410.03772","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-02T23:15:53Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c8481986700fab3a8e62100b8953376b90de4307e9cf5ec86091341c18d50d1d","abstract_canon_sha256":"7d999bd7639e46454889b753456577a9583146db129a6514095e21a425d34b75"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:19:12.791789Z","signature_b64":"xepD38K2g8IwVDCh1+0rWAc4kFI1EvjSI6JqvFc4sfcbEzFDiu5ntI4AKUw4oJ8j3kyhjCZzV6GtECRHzXclDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b0b7ef88b86529ba67b0983b06cec85b9598cd77555fb69e8cea02d516429375","last_reissued_at":"2026-07-05T09:19:12.791240Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:19:12.791240Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Precision Knowledge Editing: Enhancing Safety in Large Language Models","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Victor Bian, Xuying Li, Yasuhiro Yoshida, Yuji Kosuga, Zhuo Li","submitted_at":"2024-10-02T23:15:53Z","abstract_excerpt":"Large language models (LLMs) have demonstrated remarkable capabilities, but they also pose risks related to the generation of toxic or harmful content. This work introduces Precision Knowledge Editing (PKE), an advanced technique that builds upon existing knowledge editing methods to more effectively identify and modify toxic parameter regions within LLMs. By leveraging neuron weight tracking and activation pathway tracing, PKE achieves finer granularity in toxic content management compared to previous methods like Detoxifying Instance Neuron Modification (DINM). Our experiments demonstrate th"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.03772","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/2410.03772/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":"2410.03772","created_at":"2026-07-05T09:19:12.791321+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.03772v1","created_at":"2026-07-05T09:19:12.791321+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.03772","created_at":"2026-07-05T09:19:12.791321+00:00"},{"alias_kind":"pith_short_12","alias_value":"WC367CFYMUU3","created_at":"2026-07-05T09:19:12.791321+00:00"},{"alias_kind":"pith_short_16","alias_value":"WC367CFYMUU3UZ5Q","created_at":"2026-07-05T09:19:12.791321+00:00"},{"alias_kind":"pith_short_8","alias_value":"WC367CFY","created_at":"2026-07-05T09:19:12.791321+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.01266","citing_title":"Detoxification of Large Language Models through Output-layer Fusion with a Calibration Model","ref_index":18,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WC367CFYMUU3UZ5QTA5QNTWILO","json":"https://pith.science/pith/WC367CFYMUU3UZ5QTA5QNTWILO.json","graph_json":"https://pith.science/api/pith-number/WC367CFYMUU3UZ5QTA5QNTWILO/graph.json","events_json":"https://pith.science/api/pith-number/WC367CFYMUU3UZ5QTA5QNTWILO/events.json","paper":"https://pith.science/paper/WC367CFY"},"agent_actions":{"view_html":"https://pith.science/pith/WC367CFYMUU3UZ5QTA5QNTWILO","download_json":"https://pith.science/pith/WC367CFYMUU3UZ5QTA5QNTWILO.json","view_paper":"https://pith.science/paper/WC367CFY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.03772&json=true","fetch_graph":"https://pith.science/api/pith-number/WC367CFYMUU3UZ5QTA5QNTWILO/graph.json","fetch_events":"https://pith.science/api/pith-number/WC367CFYMUU3UZ5QTA5QNTWILO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WC367CFYMUU3UZ5QTA5QNTWILO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WC367CFYMUU3UZ5QTA5QNTWILO/action/storage_attestation","attest_author":"https://pith.science/pith/WC367CFYMUU3UZ5QTA5QNTWILO/action/author_attestation","sign_citation":"https://pith.science/pith/WC367CFYMUU3UZ5QTA5QNTWILO/action/citation_signature","submit_replication":"https://pith.science/pith/WC367CFYMUU3UZ5QTA5QNTWILO/action/replication_record"}},"created_at":"2026-07-05T09:19:12.791321+00:00","updated_at":"2026-07-05T09:19:12.791321+00:00"}