{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:GMN74I3LSV2UOARDKOK5SDPYBZ","short_pith_number":"pith:GMN74I3L","schema_version":"1.0","canonical_sha256":"331bfe236b95754702235395d90df80e695e8f606d1a2f83b632bfe1f8777186","source":{"kind":"arxiv","id":"2407.10058","version":2},"attestation_state":"computed","paper":{"title":"Learning to Refuse: Towards Mitigating Privacy Risks in LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chuanyuan Tan, Tong Zhu, Wenliang Chen, Zhenhua Liu","submitted_at":"2024-07-14T03:05:53Z","abstract_excerpt":"Large language models (LLMs) exhibit remarkable capabilities in understanding and generating natural language. However, these models can inadvertently memorize private information, posing significant privacy risks. This study addresses the challenge of enabling LLMs to protect specific individuals' private data without the need for complete retraining. We propose \\return, a Real-world pErsonal daTa UnleaRNing dataset, comprising 2,492 individuals from Wikipedia with associated QA pairs, to evaluate machine unlearning (MU) methods for protecting personal data in a realistic scenario. Additional"},"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":"2407.10058","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-14T03:05:53Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"5fcb8413aaecd1848487de65bf6104de3db7a8a31c90971ead9dcdd096b14a4e","abstract_canon_sha256":"929b723f097057f0abd5d5b58b1de52488c4e2e94e7f815e60b65e3106bac579"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:07:30.475938Z","signature_b64":"/Tcr957n4iuUzC0NArhIQwafhd9WS28PNgC9nS3BDVyRUPcAeG/k7xUGp2Op3lMpya8arMwULYMDlrIUVj8zAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"331bfe236b95754702235395d90df80e695e8f606d1a2f83b632bfe1f8777186","last_reissued_at":"2026-07-05T09:07:30.475459Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:07:30.475459Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Learning to Refuse: Towards Mitigating Privacy Risks in LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chuanyuan Tan, Tong Zhu, Wenliang Chen, Zhenhua Liu","submitted_at":"2024-07-14T03:05:53Z","abstract_excerpt":"Large language models (LLMs) exhibit remarkable capabilities in understanding and generating natural language. However, these models can inadvertently memorize private information, posing significant privacy risks. This study addresses the challenge of enabling LLMs to protect specific individuals' private data without the need for complete retraining. We propose \\return, a Real-world pErsonal daTa UnleaRNing dataset, comprising 2,492 individuals from Wikipedia with associated QA pairs, to evaluate machine unlearning (MU) methods for protecting personal data in a realistic scenario. Additional"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.10058","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/2407.10058/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":"2407.10058","created_at":"2026-07-05T09:07:30.475519+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.10058v2","created_at":"2026-07-05T09:07:30.475519+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.10058","created_at":"2026-07-05T09:07:30.475519+00:00"},{"alias_kind":"pith_short_12","alias_value":"GMN74I3LSV2U","created_at":"2026-07-05T09:07:30.475519+00:00"},{"alias_kind":"pith_short_16","alias_value":"GMN74I3LSV2UOARD","created_at":"2026-07-05T09:07:30.475519+00:00"},{"alias_kind":"pith_short_8","alias_value":"GMN74I3L","created_at":"2026-07-05T09:07:30.475519+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2503.22760","citing_title":"Malicious and Unintentional Disclosure Risks in Large Language Models for Code Generation","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21539","citing_title":"DualOptim+: Bridging Shared and Decoupled Optimizer States for Better Machine Unlearning in Large Language Models","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08765","citing_title":"Unlearners Can Lie: Evaluating and Improving Honesty in LLM Unlearning","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2604.14644","citing_title":"CURaTE: Continual Unlearning in Real Time with Ensured Preservation of LLM Knowledge","ref_index":2,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GMN74I3LSV2UOARDKOK5SDPYBZ","json":"https://pith.science/pith/GMN74I3LSV2UOARDKOK5SDPYBZ.json","graph_json":"https://pith.science/api/pith-number/GMN74I3LSV2UOARDKOK5SDPYBZ/graph.json","events_json":"https://pith.science/api/pith-number/GMN74I3LSV2UOARDKOK5SDPYBZ/events.json","paper":"https://pith.science/paper/GMN74I3L"},"agent_actions":{"view_html":"https://pith.science/pith/GMN74I3LSV2UOARDKOK5SDPYBZ","download_json":"https://pith.science/pith/GMN74I3LSV2UOARDKOK5SDPYBZ.json","view_paper":"https://pith.science/paper/GMN74I3L","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.10058&json=true","fetch_graph":"https://pith.science/api/pith-number/GMN74I3LSV2UOARDKOK5SDPYBZ/graph.json","fetch_events":"https://pith.science/api/pith-number/GMN74I3LSV2UOARDKOK5SDPYBZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GMN74I3LSV2UOARDKOK5SDPYBZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GMN74I3LSV2UOARDKOK5SDPYBZ/action/storage_attestation","attest_author":"https://pith.science/pith/GMN74I3LSV2UOARDKOK5SDPYBZ/action/author_attestation","sign_citation":"https://pith.science/pith/GMN74I3LSV2UOARDKOK5SDPYBZ/action/citation_signature","submit_replication":"https://pith.science/pith/GMN74I3LSV2UOARDKOK5SDPYBZ/action/replication_record"}},"created_at":"2026-07-05T09:07:30.475519+00:00","updated_at":"2026-07-05T09:07:30.475519+00:00"}