{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:NFZUJRGFD4FF5VAOP56TFBFWUG","short_pith_number":"pith:NFZUJRGF","schema_version":"1.0","canonical_sha256":"697344c4c51f0a5ed40e7f7d3284b6a18a7649eee1abb197bfb1eaa5f530df67","source":{"kind":"arxiv","id":"2407.06460","version":2},"attestation_state":"computed","paper":{"title":"MUSE: Machine Unlearning Six-Way Evaluation for Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Ari Holtzman, Chiyuan Zhang, Daogao Liu, Jaechan Lee, Jieyu Zhao, Luke Zettlemoyer, Noah A. Smith, Sadhika Malladi, Weijia Shi, Yangsibo Huang","submitted_at":"2024-07-08T23:47:29Z","abstract_excerpt":"Language models (LMs) are trained on vast amounts of text data, which may include private and copyrighted content. Data owners may request the removal of their data from a trained model due to privacy or copyright concerns. However, exactly unlearning only these datapoints (i.e., retraining with the data removed) is intractable in modern-day models. This has led to the development of many approximate unlearning algorithms. The evaluation of the efficacy of these algorithms has traditionally been narrow in scope, failing to precisely quantify the success and practicality of the algorithm from t"},"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":"2407.06460","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-08T23:47:29Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"658c0965bbd913b705d5bcfa063a8b8af00cdfce01c70e9e888228110e0f2779","abstract_canon_sha256":"490534ecd06b5de84f03dd4408cac75d21cf1eee820b293e81c0916ba5d0c735"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:43:30.308418Z","signature_b64":"chMSO0l9AQ8JkzRaTqgy1fAb6tZHW814Xsux/SJ3c2cRQaw7z/F/GEa4oJo9cTtyG7x6/sgHi09WBV6mxIkAAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"697344c4c51f0a5ed40e7f7d3284b6a18a7649eee1abb197bfb1eaa5f530df67","last_reissued_at":"2026-07-05T08:43:30.307965Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:43:30.307965Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MUSE: Machine Unlearning Six-Way Evaluation for Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Ari Holtzman, Chiyuan Zhang, Daogao Liu, Jaechan Lee, Jieyu Zhao, Luke Zettlemoyer, Noah A. Smith, Sadhika Malladi, Weijia Shi, Yangsibo Huang","submitted_at":"2024-07-08T23:47:29Z","abstract_excerpt":"Language models (LMs) are trained on vast amounts of text data, which may include private and copyrighted content. Data owners may request the removal of their data from a trained model due to privacy or copyright concerns. However, exactly unlearning only these datapoints (i.e., retraining with the data removed) is intractable in modern-day models. This has led to the development of many approximate unlearning algorithms. The evaluation of the efficacy of these algorithms has traditionally been narrow in scope, failing to precisely quantify the success and practicality of the algorithm from t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.06460","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/2407.06460/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":"2407.06460","created_at":"2026-07-05T08:43:30.308021+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.06460v2","created_at":"2026-07-05T08:43:30.308021+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.06460","created_at":"2026-07-05T08:43:30.308021+00:00"},{"alias_kind":"pith_short_12","alias_value":"NFZUJRGFD4FF","created_at":"2026-07-05T08:43:30.308021+00:00"},{"alias_kind":"pith_short_16","alias_value":"NFZUJRGFD4FF5VAO","created_at":"2026-07-05T08:43:30.308021+00:00"},{"alias_kind":"pith_short_8","alias_value":"NFZUJRGF","created_at":"2026-07-05T08:43:30.308021+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":23,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.18473","citing_title":"PreUnlearn: Auditing Collateral Knowledge Damage Before Large Language Model Unlearning","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2606.17168","citing_title":"RepSelect: Robust LLM Unlearning via Representation Selectivity","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2606.12841","citing_title":"TimeROME-DLM: Temporal Causal Tracing and Low-Rank Inference-Time Knowledge Editing for Masked Diffusion Language Models","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2606.25198","citing_title":"Heuresis: Search Strategies for Autonomous AI Research Agents Across Quality, Diversity and Novelty","ref_index":57,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07141","citing_title":"REMEDI: A Benchmark for Retention and Unlearning Evaluation in Multi-label Clinical Disease Inference","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2606.05029","citing_title":"Validity Threats for Foundation Model Research","ref_index":88,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02920","citing_title":"Fast Unlearning at Scale via Margin Self-Correction","ref_index":44,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01129","citing_title":"Revisiting Privacy Leakage in Machine Unlearning: Membership Inference Beyond the Forgotten Set","ref_index":73,"is_internal_anchor":false},{"citing_arxiv_id":"2605.30514","citing_title":"MAAT: Multi-phase Adapter-Aware Targeted Unlearning","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2605.30919","citing_title":"De-attribute to Forget for LLM Unlearning","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20915","citing_title":"Calibration vs Decision Making: Revisiting the Reliability Paradox in Unlearned Language Models","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2506.20941","citing_title":"Revisiting the Past: Data Unlearning with Model State History","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2510.00761","citing_title":"Downgrade to Upgrade: Optimizer Simplification Enhances Robustness in LLM Unlearning","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14404","citing_title":"Knowledge Beyond Language: Bridging the Gap in Multilingual Machine Unlearning Evaluation","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2604.03114","citing_title":"Can VLMs Truly Forget? Benchmarking Training-Free Visual Concept Unlearning","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11685","citing_title":"Robust LLM Unlearning Against Relearning Attacks: The Minor Components in Representations Matter","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05938","citing_title":"ICU-Bench:Benchmarking Continual Unlearning in Multimodal Large Language Models","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02206","citing_title":"Metric Unreliability in Multimodal Machine Unlearning: A Systematic Analysis and Principled Unified Score","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01129","citing_title":"Revisiting Privacy Leakage in Machine Unlearning: Membership Inference Beyond the Forgotten Set","ref_index":73,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01735","citing_title":"Less is More: Geometric Unlearning for LLMs with Minimal Data Disclosure","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07962","citing_title":"Is your algorithm unlearning or untraining?","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02206","citing_title":"Metric Unreliability in Multimodal Machine Unlearning: A Systematic Analysis and Principled Unified Score","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2604.13438","citing_title":"WIN-U: Woodbury-Informed Newton-Unlearning as a retain-free Machine Unlearning Framework","ref_index":17,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NFZUJRGFD4FF5VAOP56TFBFWUG","json":"https://pith.science/pith/NFZUJRGFD4FF5VAOP56TFBFWUG.json","graph_json":"https://pith.science/api/pith-number/NFZUJRGFD4FF5VAOP56TFBFWUG/graph.json","events_json":"https://pith.science/api/pith-number/NFZUJRGFD4FF5VAOP56TFBFWUG/events.json","paper":"https://pith.science/paper/NFZUJRGF"},"agent_actions":{"view_html":"https://pith.science/pith/NFZUJRGFD4FF5VAOP56TFBFWUG","download_json":"https://pith.science/pith/NFZUJRGFD4FF5VAOP56TFBFWUG.json","view_paper":"https://pith.science/paper/NFZUJRGF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.06460&json=true","fetch_graph":"https://pith.science/api/pith-number/NFZUJRGFD4FF5VAOP56TFBFWUG/graph.json","fetch_events":"https://pith.science/api/pith-number/NFZUJRGFD4FF5VAOP56TFBFWUG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NFZUJRGFD4FF5VAOP56TFBFWUG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NFZUJRGFD4FF5VAOP56TFBFWUG/action/storage_attestation","attest_author":"https://pith.science/pith/NFZUJRGFD4FF5VAOP56TFBFWUG/action/author_attestation","sign_citation":"https://pith.science/pith/NFZUJRGFD4FF5VAOP56TFBFWUG/action/citation_signature","submit_replication":"https://pith.science/pith/NFZUJRGFD4FF5VAOP56TFBFWUG/action/replication_record"}},"created_at":"2026-07-05T08:43:30.308021+00:00","updated_at":"2026-07-05T08:43:30.308021+00:00"}