{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:DUUMFKS3MXJP5RFCTPHS2DGPFB","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"cdfb1d4bacb891d0760250c6834fbfe6c4d422f9bedf3a0b27b9b061f7ce4a91","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-30T10:07:52Z","title_canon_sha256":"e7c3bc6d1acf0a5f8de4f9ac48ec4995b17e18638e037301dc4a98f76c2ef66e"},"schema_version":"1.0","source":{"id":"2505.24428","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.24428","created_at":"2026-07-05T11:12:49Z"},{"alias_kind":"arxiv_version","alias_value":"2505.24428v1","created_at":"2026-07-05T11:12:49Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.24428","created_at":"2026-07-05T11:12:49Z"},{"alias_kind":"pith_short_12","alias_value":"DUUMFKS3MXJP","created_at":"2026-07-05T11:12:49Z"},{"alias_kind":"pith_short_16","alias_value":"DUUMFKS3MXJP5RFC","created_at":"2026-07-05T11:12:49Z"},{"alias_kind":"pith_short_8","alias_value":"DUUMFKS3","created_at":"2026-07-05T11:12:49Z"}],"graph_snapshots":[{"event_id":"sha256:a57077ab0c6f139545035bab89fe6f2c920344e1bc83906e56999cb40683a01b","target":"graph","created_at":"2026-07-05T11:12:49Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2505.24428/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) store vast amounts of information, making them powerful yet raising privacy and safety concerns when selective knowledge removal is required. Existing unlearning strategies, ranging from gradient-based fine-tuning and model editing to sparse autoencoder (SAE) steering, either lack interpretability or fail to provide a robust defense against adversarial prompts. We propose SAE-Guided Subspace Projection Unlearning (SSPU), a novel framework that leverages SAE features to drive targeted updates in the model's parameter space, enabling precise, interpretable, and robus","authors_text":"Benyou Wang, Difan Zou, Xu Wang, Yan Hu, Zihao Li","cross_cats":["cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-30T10:07:52Z","title":"Model Unlearning via Sparse Autoencoder Subspace Guided Projections"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.24428","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:6d2ca7bc74d22c1f017c88f077426ef4810fce8135d3d83321838ccf246f948d","target":"record","created_at":"2026-07-05T11:12:49Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"cdfb1d4bacb891d0760250c6834fbfe6c4d422f9bedf3a0b27b9b061f7ce4a91","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-30T10:07:52Z","title_canon_sha256":"e7c3bc6d1acf0a5f8de4f9ac48ec4995b17e18638e037301dc4a98f76c2ef66e"},"schema_version":"1.0","source":{"id":"2505.24428","kind":"arxiv","version":1}},"canonical_sha256":"1d28c2aa5b65d2fec4a29bcf2d0ccf284a2b377cd9b0156afe7e5eb7ff22656b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1d28c2aa5b65d2fec4a29bcf2d0ccf284a2b377cd9b0156afe7e5eb7ff22656b","first_computed_at":"2026-07-05T11:12:49.554514Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:12:49.554514Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"R6bumi397xxUENFd1lrRYAAmHCSZ+vWNz7MDzNYZN4ssM0XS/1eozeipDRDO3fMuQ/RxsgN/OYY/minEYBIkAQ==","signature_status":"signed_v1","signed_at":"2026-07-05T11:12:49.555015Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.24428","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6d2ca7bc74d22c1f017c88f077426ef4810fce8135d3d83321838ccf246f948d","sha256:a57077ab0c6f139545035bab89fe6f2c920344e1bc83906e56999cb40683a01b"],"state_sha256":"c1d919e904fc66d0891b5024c07ed01cbc37f2d7c7488e1f59743f607dc154dd"}