{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:PQRPIRV7LCPVV7RJKTX5EA35PO","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":"63baf00f9ffb5276a044a4afd373a20a2e5e1da8fdbfc54d90bfde73f7b9166d","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-03T18:59:01Z","title_canon_sha256":"bb2b3d3ccc21ede7d8ad21fab5c8f58c49993057ce236f118f821ba40cb19d7a"},"schema_version":"1.0","source":{"id":"2502.01633","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.01633","created_at":"2026-07-05T11:26:47Z"},{"alias_kind":"arxiv_version","alias_value":"2502.01633v2","created_at":"2026-07-05T11:26:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.01633","created_at":"2026-07-05T11:26:47Z"},{"alias_kind":"pith_short_12","alias_value":"PQRPIRV7LCPV","created_at":"2026-07-05T11:26:47Z"},{"alias_kind":"pith_short_16","alias_value":"PQRPIRV7LCPVV7RJ","created_at":"2026-07-05T11:26:47Z"},{"alias_kind":"pith_short_8","alias_value":"PQRPIRV7","created_at":"2026-07-05T11:26:47Z"}],"graph_snapshots":[{"event_id":"sha256:22b13596ee5e1e4536fc8807264286d422a21545ad2056eb2b9cbd7dbd41a199","target":"graph","created_at":"2026-07-05T11:26:47Z","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/2502.01633/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"As large language models (LLMs) are becoming more capable and widespread, the study of their failure cases is becoming increasingly important. Recent advances in standardizing, measuring, and scaling test-time compute suggest new methodologies for optimizing models to achieve high performance on hard tasks. In this paper, we apply these advances to the task of model jailbreaking: eliciting harmful responses from aligned LLMs. We develop an adversarial reasoning approach to automatic jailbreaking that leverages a loss signal to guide the test-time compute, achieving SOTA attack success rates ag","authors_text":"Amin Karbasi, George Pappas, Hamed Hassani, Mahdi Sabbaghi, Paul Kassianik, Yaron Singer","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-03T18:59:01Z","title":"Adversarial Reasoning at Jailbreaking Time"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.01633","kind":"arxiv","version":2},"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:a96ce78fa4d3fcb77dbbc83398269da37b0b6c512dba06923b641f1043f4b645","target":"record","created_at":"2026-07-05T11:26:47Z","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":"63baf00f9ffb5276a044a4afd373a20a2e5e1da8fdbfc54d90bfde73f7b9166d","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-03T18:59:01Z","title_canon_sha256":"bb2b3d3ccc21ede7d8ad21fab5c8f58c49993057ce236f118f821ba40cb19d7a"},"schema_version":"1.0","source":{"id":"2502.01633","kind":"arxiv","version":2}},"canonical_sha256":"7c22f446bf589f5afe2954efd2037d7ba5c7b7abbe541ca9787a41d6ecf07dee","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7c22f446bf589f5afe2954efd2037d7ba5c7b7abbe541ca9787a41d6ecf07dee","first_computed_at":"2026-07-05T11:26:47.393383Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:26:47.393383Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"GdyQvb4oWLiZorRQxlakMvTM1nR73uW7VU/jFN14P5RtjxEZeboGy7qVIIqvZIAhFiK8m9lDR+idaEsKK8JPCA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:26:47.393864Z","signed_message":"canonical_sha256_bytes"},"source_id":"2502.01633","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a96ce78fa4d3fcb77dbbc83398269da37b0b6c512dba06923b641f1043f4b645","sha256:22b13596ee5e1e4536fc8807264286d422a21545ad2056eb2b9cbd7dbd41a199"],"state_sha256":"51efa2ed0440d511d6a96c6f596fbb95c376ab198aa48df116819a044f42eba6"}