{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:6XNFRFD2EQ3PXH3BY3KGCSVJJL","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":"a43686c9539b199548daff086f62559ea5aab6b08bcfc294efc9b08c42269313","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2025-08-27T09:06:28Z","title_canon_sha256":"b0da9e974128f42e649c73e03bbe2344ceaf512e8aeeb66af155ab845bee3f0e"},"schema_version":"1.0","source":{"id":"2508.19697","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.19697","created_at":"2026-07-05T12:00:16Z"},{"alias_kind":"arxiv_version","alias_value":"2508.19697v1","created_at":"2026-07-05T12:00:16Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.19697","created_at":"2026-07-05T12:00:16Z"},{"alias_kind":"pith_short_12","alias_value":"6XNFRFD2EQ3P","created_at":"2026-07-05T12:00:16Z"},{"alias_kind":"pith_short_16","alias_value":"6XNFRFD2EQ3PXH3B","created_at":"2026-07-05T12:00:16Z"},{"alias_kind":"pith_short_8","alias_value":"6XNFRFD2","created_at":"2026-07-05T12:00:16Z"}],"graph_snapshots":[{"event_id":"sha256:ffde783d419011b08ab98bf368d15fd459ca84a73f60924e7133b85ab10d78ad","target":"graph","created_at":"2026-07-05T12:00:16Z","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/2508.19697/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Current safety alignment for large language models(LLMs) continues to present vulnerabilities, given that adversarial prompting can effectively bypass their safety measures.Our investigation shows that these safety mechanisms predominantly depend on a limited subset of attention heads: removing or ablating these heads can severely compromise model safety. To identify and evaluate these safety-critical components, we introduce RDSHA, a targeted ablation method that leverages the model's refusal direction to pinpoint attention heads mostly responsible for safety behaviors. Further analysis shows","authors_text":"Chao Huang, Chuang Zhang, Juewei Yue, Quangang Li, Tingwen Liu, Zefeng Zhang","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2025-08-27T09:06:28Z","title":"Safety Alignment Should Be Made More Than Just A Few Attention Heads"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.19697","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":"sha256:41421a9879130415e1062c9fbdb7a0bfb57df7b76aa1e42eece8b7d95d230464","target":"record","created_at":"2026-07-05T12:00:16Z","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":"a43686c9539b199548daff086f62559ea5aab6b08bcfc294efc9b08c42269313","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2025-08-27T09:06:28Z","title_canon_sha256":"b0da9e974128f42e649c73e03bbe2344ceaf512e8aeeb66af155ab845bee3f0e"},"schema_version":"1.0","source":{"id":"2508.19697","kind":"arxiv","version":1}},"canonical_sha256":"f5da58947a2436fb9f61c6d4614aa94aece857d55956410e15c2fe3d79ef27f6","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f5da58947a2436fb9f61c6d4614aa94aece857d55956410e15c2fe3d79ef27f6","first_computed_at":"2026-07-05T12:00:16.624580Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T12:00:16.624580Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Ur150CSFrukjcGtdR4EIagcD2hup/NencMnKeTE8fUQ9lUHw2VzyyCbvMUbIOM6oY5QugD5BmtVamAEBIKEvDA==","signature_status":"signed_v1","signed_at":"2026-07-05T12:00:16.625034Z","signed_message":"canonical_sha256_bytes"},"source_id":"2508.19697","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:41421a9879130415e1062c9fbdb7a0bfb57df7b76aa1e42eece8b7d95d230464","sha256:ffde783d419011b08ab98bf368d15fd459ca84a73f60924e7133b85ab10d78ad"],"state_sha256":"dc26b13c84bcff8ac0cb22899e82e4950d35d6af7ba0f9c7858a704f8d48fa1b"}