{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:NXOEHGX4ODJRUXKX23MJ54CXWR","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":"23a9e458d55e85cfdd3d1830062f32601294ae79bdb8282af27531f29ff76fae","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-18T09:46:25Z","title_canon_sha256":"603bcf7a2e0afa0fb7f07f102e5acdc40a5aa5928e7e26c10e073b16178bcce9"},"schema_version":"1.0","source":{"id":"2404.12038","kind":"arxiv","version":5}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.12038","created_at":"2026-07-05T09:42:24Z"},{"alias_kind":"arxiv_version","alias_value":"2404.12038v5","created_at":"2026-07-05T09:42:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.12038","created_at":"2026-07-05T09:42:24Z"},{"alias_kind":"pith_short_12","alias_value":"NXOEHGX4ODJR","created_at":"2026-07-05T09:42:24Z"},{"alias_kind":"pith_short_16","alias_value":"NXOEHGX4ODJRUXKX","created_at":"2026-07-05T09:42:24Z"},{"alias_kind":"pith_short_8","alias_value":"NXOEHGX4","created_at":"2026-07-05T09:42:24Z"}],"graph_snapshots":[{"event_id":"sha256:3ee766f0d9b5529dc207ea9ee47d8a5a4158c5daf8e1665e8a7861569b1f5cdc","target":"graph","created_at":"2026-07-05T09:42:24Z","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/2404.12038/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Despite careful safety alignment, current large language models (LLMs) remain vulnerable to various attacks. To further unveil the safety risks of LLMs, we introduce a Safety Concept Activation Vector (SCAV) framework, which effectively guides the attacks by accurately interpreting LLMs' safety mechanisms. We then develop an SCAV-guided attack method that can generate both attack prompts and embedding-level attacks with automatically selected perturbation hyperparameters. Both automatic and human evaluations demonstrate that our attack method significantly improves the attack success rate and ","authors_text":"Changyu Chen, Ruixuan Huang, Xiting Wang, Zhihao Xu","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-18T09:46:25Z","title":"Uncovering Safety Risks of Large Language Models through Concept Activation Vector"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.12038","kind":"arxiv","version":5},"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":"sha256:25db4c452679e0a3a1ff5a69c765f01029f9152519739cb6263f26039dd79f05","target":"record","created_at":"2026-07-05T09:42:24Z","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":"23a9e458d55e85cfdd3d1830062f32601294ae79bdb8282af27531f29ff76fae","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-18T09:46:25Z","title_canon_sha256":"603bcf7a2e0afa0fb7f07f102e5acdc40a5aa5928e7e26c10e073b16178bcce9"},"schema_version":"1.0","source":{"id":"2404.12038","kind":"arxiv","version":5}},"canonical_sha256":"6ddc439afc70d31a5d57d6d89ef057b461b16ec73b4b5d08ee6f4195ff751992","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"6ddc439afc70d31a5d57d6d89ef057b461b16ec73b4b5d08ee6f4195ff751992","first_computed_at":"2026-07-05T09:42:24.692264Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:42:24.692264Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"obD2CpoCdJJvT7GQkAD7qrP0SfgppXa9GcyM2iJMMsZcXDszcu2bmXJN6byT+kzys2s93v9DqxDmxpah9aWjCQ==","signature_status":"signed_v1","signed_at":"2026-07-05T09:42:24.692782Z","signed_message":"canonical_sha256_bytes"},"source_id":"2404.12038","source_kind":"arxiv","source_version":5}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:25db4c452679e0a3a1ff5a69c765f01029f9152519739cb6263f26039dd79f05","sha256:3ee766f0d9b5529dc207ea9ee47d8a5a4158c5daf8e1665e8a7861569b1f5cdc"],"state_sha256":"131ee5432c31399a0ac51141c850332874a79eae2d00fc40e0f2e3970ca37561"}