{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:FJLBWM64XKLFNTVM7AEHIJHXBQ","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":"6de302ea871995bf834c85af4825781be8c66effaa060a9568a0702876c46802","cross_cats_sorted":["cs.CL","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-04-02T09:47:01Z","title_canon_sha256":"0f77129d90000c962e922e213a51ec1bfccee082ea31931ec4aec63f52dceaad"},"schema_version":"1.0","source":{"id":"2504.01550","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2504.01550","created_at":"2026-07-05T11:37:17Z"},{"alias_kind":"arxiv_version","alias_value":"2504.01550v3","created_at":"2026-07-05T11:37:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.01550","created_at":"2026-07-05T11:37:17Z"},{"alias_kind":"pith_short_12","alias_value":"FJLBWM64XKLF","created_at":"2026-07-05T11:37:17Z"},{"alias_kind":"pith_short_16","alias_value":"FJLBWM64XKLFNTVM","created_at":"2026-07-05T11:37:17Z"},{"alias_kind":"pith_short_8","alias_value":"FJLBWM64","created_at":"2026-07-05T11:37:17Z"}],"graph_snapshots":[{"event_id":"sha256:8d4586ca4f9a278f078f18c1bfd11d1825e6726174451afc0f5359b734ffea44","target":"graph","created_at":"2026-07-05T11:37:17Z","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/2504.01550/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Models (LLMs) have emerged as powerful tools, but their inherent safety risks - ranging from harmful content generation to broader societal harms - pose significant challenges. These risks can be amplified by the recent adversarial attacks, fine-tuning vulnerabilities, and the increasing deployment of LLMs in high-stakes environments. Existing safety-enhancing techniques, such as fine-tuning with human feedback or adversarial training, are still vulnerable as they address specific threats and often fail to generalize across unseen attacks, or require manual system-level defenses","authors_text":"Alvin Wan, Ashkan Yousefpour, Harrison Ngan, Jonghyun Choi, Ryan S. Kwon, Seungbeen Lee, Seungju Han, Taeheon Kim, Wonje Jeung, Youngjae Yu","cross_cats":["cs.CL","cs.CR"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-04-02T09:47:01Z","title":"Representation Bending for Large Language Model Safety"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.01550","kind":"arxiv","version":3},"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":"sha256:1c6c5813d75266fbe418808fd32299596ce92bab22f35f908454605bc1c9fc8b","target":"record","created_at":"2026-07-05T11:37:17Z","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":"6de302ea871995bf834c85af4825781be8c66effaa060a9568a0702876c46802","cross_cats_sorted":["cs.CL","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-04-02T09:47:01Z","title_canon_sha256":"0f77129d90000c962e922e213a51ec1bfccee082ea31931ec4aec63f52dceaad"},"schema_version":"1.0","source":{"id":"2504.01550","kind":"arxiv","version":3}},"canonical_sha256":"2a561b33dcba9656ceacf8087424f70c1004e743b3b324c4f0d6e71117bb6b77","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2a561b33dcba9656ceacf8087424f70c1004e743b3b324c4f0d6e71117bb6b77","first_computed_at":"2026-07-05T11:37:17.599283Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:37:17.599283Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"JFN6beeiEae1+7X7iYb+iumcT2Gez6TkJEYM68V5C3HYYN/GZRpKdx66APbaaOTBa/oVARfQfsR8RLrDmnmsCA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:37:17.599873Z","signed_message":"canonical_sha256_bytes"},"source_id":"2504.01550","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1c6c5813d75266fbe418808fd32299596ce92bab22f35f908454605bc1c9fc8b","sha256:8d4586ca4f9a278f078f18c1bfd11d1825e6726174451afc0f5359b734ffea44"],"state_sha256":"193dbc4d666f1de7703cfdeeced84147b8b41d6a4484bcd576b832886e983d9c"}