{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:QJ5G6AOYIYCALI5NNYBWLF2QRB","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":"8cec4f3854658bc079022f825abd60c62cc5198d3db392d4445d0caef7c598be","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2024-12-10T01:30:32Z","title_canon_sha256":"e8611ea233b83230901da795a21219a6f13a0674194aaeea5b083e9a3bed5539"},"schema_version":"1.0","source":{"id":"2412.07097","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2412.07097","created_at":"2026-07-05T09:47:03Z"},{"alias_kind":"arxiv_version","alias_value":"2412.07097v1","created_at":"2026-07-05T09:47:03Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.07097","created_at":"2026-07-05T09:47:03Z"},{"alias_kind":"pith_short_12","alias_value":"QJ5G6AOYIYCA","created_at":"2026-07-05T09:47:03Z"},{"alias_kind":"pith_short_16","alias_value":"QJ5G6AOYIYCALI5N","created_at":"2026-07-05T09:47:03Z"},{"alias_kind":"pith_short_8","alias_value":"QJ5G6AOY","created_at":"2026-07-05T09:47:03Z"}],"graph_snapshots":[{"event_id":"sha256:591740733515dd1470854cc97a1e604a087e45655557464b9a94c0ff4dcd0147","target":"graph","created_at":"2026-07-05T09:47:03Z","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/2412.07097/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Stakeholders -- from model developers to policymakers -- seek to minimize the dual-use risks of large language models (LLMs). An open challenge to this goal is whether technical safeguards can impede the misuse of LLMs, even when models are customizable via fine-tuning or when model weights are fully open. In response, several recent studies have proposed methods to produce durable LLM safeguards for open-weight LLMs that can withstand adversarial modifications of the model's weights via fine-tuning. This holds the promise of raising adversaries' costs even under strong threat models where adv","authors_text":"Boyi Wei, Luxi He, Matthew Jagielski, Milad Nasr, Nicholas Carlini, Peter Henderson, Prateek Mittal, Tinghao Xie, Xiangyu Qi, Yangsibo Huang","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2024-12-10T01:30:32Z","title":"On Evaluating the Durability of Safeguards for Open-Weight LLMs"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.07097","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":"sha256:907c68ea84cc99c459fb761b647a78e47b12e6c8e3bab9d5c628deaf94d7039e","target":"record","created_at":"2026-07-05T09:47:03Z","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":"8cec4f3854658bc079022f825abd60c62cc5198d3db392d4445d0caef7c598be","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2024-12-10T01:30:32Z","title_canon_sha256":"e8611ea233b83230901da795a21219a6f13a0674194aaeea5b083e9a3bed5539"},"schema_version":"1.0","source":{"id":"2412.07097","kind":"arxiv","version":1}},"canonical_sha256":"827a6f01d8460405a3ad6e0365975088777e5a0f5e713aa4e68cf63aa53d7555","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"827a6f01d8460405a3ad6e0365975088777e5a0f5e713aa4e68cf63aa53d7555","first_computed_at":"2026-07-05T09:47:03.539898Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:47:03.539898Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"lCoS2TzZtczMEDfUdj2owHWX40Y1SkbSOtMEEiRTDPdYlfEyCMygg7DPk+gpbNRQccbx++ZgJIIrurznxGzjAw==","signature_status":"signed_v1","signed_at":"2026-07-05T09:47:03.540382Z","signed_message":"canonical_sha256_bytes"},"source_id":"2412.07097","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:907c68ea84cc99c459fb761b647a78e47b12e6c8e3bab9d5c628deaf94d7039e","sha256:591740733515dd1470854cc97a1e604a087e45655557464b9a94c0ff4dcd0147"],"state_sha256":"757787aaf2f9f56f55702a8930345b4ad3b71c4cefdb7a651d2336856592f196"}