{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:BR6IIYKLCTLW36TO2QEFOBVHTI","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":"87ea1f924391097357bd775ecd7194c91290b736455b16857c8da94f69f74736","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2025-06-07T08:19:01Z","title_canon_sha256":"db86fe356bd8d592fab109186f9d7867f399a2a65a6afc2f013b3c1ae7f8397a"},"schema_version":"1.0","source":{"id":"2506.10020","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.10020","created_at":"2026-07-05T11:20:38Z"},{"alias_kind":"arxiv_version","alias_value":"2506.10020v1","created_at":"2026-07-05T11:20:38Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.10020","created_at":"2026-07-05T11:20:38Z"},{"alias_kind":"pith_short_12","alias_value":"BR6IIYKLCTLW","created_at":"2026-07-05T11:20:38Z"},{"alias_kind":"pith_short_16","alias_value":"BR6IIYKLCTLW36TO","created_at":"2026-07-05T11:20:38Z"},{"alias_kind":"pith_short_8","alias_value":"BR6IIYKL","created_at":"2026-07-05T11:20:38Z"}],"graph_snapshots":[{"event_id":"sha256:aa0f54193f853fcbe1f070aa9822625efceb88e2062dc4a7cc773f2027812c87","target":"graph","created_at":"2026-07-05T11:20:38Z","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/2506.10020/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Safely aligning large language models (LLMs) often demands extensive human-labeled preference data, a process that's both costly and time-consuming. While synthetic data offers a promising alternative, current methods frequently rely on complex iterative prompting or auxiliary models. To address this, we introduce Refusal-Aware Adaptive Injection (RAAI), a straightforward, training-free, and model-agnostic framework that repurposes LLM attack techniques. RAAI works by detecting internal refusal signals and adaptively injecting predefined phrases to elicit harmful, yet fluent, completions. Our ","authors_text":"Hyunbin Jin, Kyubyung Chae, Taesup Kim","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2025-06-07T08:19:01Z","title":"From Threat to Tool: Leveraging Refusal-Aware Injection Attacks for Safety Alignment"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.10020","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:b3d3d2bb3efa76e4f63c4292aea1c49ce4fd4d14ba74c4781cd4973291598ba2","target":"record","created_at":"2026-07-05T11:20:38Z","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":"87ea1f924391097357bd775ecd7194c91290b736455b16857c8da94f69f74736","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2025-06-07T08:19:01Z","title_canon_sha256":"db86fe356bd8d592fab109186f9d7867f399a2a65a6afc2f013b3c1ae7f8397a"},"schema_version":"1.0","source":{"id":"2506.10020","kind":"arxiv","version":1}},"canonical_sha256":"0c7c84614b14d76dfa6ed4085706a79a3347fcc8bc68856ba321eb4bc89178bf","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0c7c84614b14d76dfa6ed4085706a79a3347fcc8bc68856ba321eb4bc89178bf","first_computed_at":"2026-07-05T11:20:38.608375Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:20:38.608375Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Rf1MTNPlo+USR1d3beadkqDhZMSfb91PZcxZeoW2zWVWIcdvzQ1zxS5sGUJPiiBLd96cCLXEbNmFiIstoN/TCA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:20:38.608827Z","signed_message":"canonical_sha256_bytes"},"source_id":"2506.10020","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b3d3d2bb3efa76e4f63c4292aea1c49ce4fd4d14ba74c4781cd4973291598ba2","sha256:aa0f54193f853fcbe1f070aa9822625efceb88e2062dc4a7cc773f2027812c87"],"state_sha256":"c12e5fdc16ec129d9b74a89f8b48f0ea8cad6166143403da9fc7087540e333c9"}