{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:YIB7NY4OG2KBXGXRA3GT7GDY7G","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":"cee2b5d204ac742b17aac8449cbf00a526888c52b0fb20c81c92f01d8466dbf9","cross_cats_sorted":["cs.AI","cs.CR","cs.LG"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-02T15:15:38Z","title_canon_sha256":"8605b3c9145f5e26939091e64091cbc772f3d4b14f09540865f8a0086911ce94"},"schema_version":"1.0","source":{"id":"2501.02018","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.02018","created_at":"2026-07-05T09:57:00Z"},{"alias_kind":"arxiv_version","alias_value":"2501.02018v1","created_at":"2026-07-05T09:57:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.02018","created_at":"2026-07-05T09:57:00Z"},{"alias_kind":"pith_short_12","alias_value":"YIB7NY4OG2KB","created_at":"2026-07-05T09:57:00Z"},{"alias_kind":"pith_short_16","alias_value":"YIB7NY4OG2KBXGXR","created_at":"2026-07-05T09:57:00Z"},{"alias_kind":"pith_short_8","alias_value":"YIB7NY4O","created_at":"2026-07-05T09:57:00Z"}],"graph_snapshots":[{"event_id":"sha256:cf6ad3a3d3149a577aff0febce57c5511d5fb5bd89f207ad2a10982104c2bcc3","target":"graph","created_at":"2026-07-05T09:57:00Z","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/2501.02018/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Models (LLMs) have been shown to be susceptible to jailbreak attacks, or adversarial attacks used to illicit high risk behavior from a model. Jailbreaks have been exploited by cybercriminals and blackhat actors to cause significant harm, highlighting the critical need to safeguard widely-deployed models. Safeguarding approaches, which include fine-tuning models or having LLMs \"self-reflect\", may lengthen the inference time of a model, incur a computational penalty, reduce the semantic fluency of an output, and restrict ``normal'' model behavior. Importantly, these Safety-Perform","authors_text":"Andrew Bell, Joao Fonseca, Julia Stoyanovich","cross_cats":["cs.AI","cs.CR","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-02T15:15:38Z","title":"Safeguarding Large Language Models in Real-time with Tunable Safety-Performance Trade-offs"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.02018","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":"sha256:1598121bb3bc515bcc71f9f43d90f6d598cdb293602d87275f4d691c208ec0a1","target":"record","created_at":"2026-07-05T09:57:00Z","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":"cee2b5d204ac742b17aac8449cbf00a526888c52b0fb20c81c92f01d8466dbf9","cross_cats_sorted":["cs.AI","cs.CR","cs.LG"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-02T15:15:38Z","title_canon_sha256":"8605b3c9145f5e26939091e64091cbc772f3d4b14f09540865f8a0086911ce94"},"schema_version":"1.0","source":{"id":"2501.02018","kind":"arxiv","version":1}},"canonical_sha256":"c203f6e38e36941b9af106cd3f9878f9b89c414a0454a3594580c9d56518038a","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c203f6e38e36941b9af106cd3f9878f9b89c414a0454a3594580c9d56518038a","first_computed_at":"2026-07-05T09:57:00.374375Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:57:00.374375Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"0vrsW9Ipft8yS8vwPHFeCwKtzjuHQJ55FnbEZgGWjP6WE5ePZBu08RKOvt7OmvFPcjbnjfxQIaosNDsI2FXNBg==","signature_status":"signed_v1","signed_at":"2026-07-05T09:57:00.374751Z","signed_message":"canonical_sha256_bytes"},"source_id":"2501.02018","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1598121bb3bc515bcc71f9f43d90f6d598cdb293602d87275f4d691c208ec0a1","sha256:cf6ad3a3d3149a577aff0febce57c5511d5fb5bd89f207ad2a10982104c2bcc3"],"state_sha256":"354055e04ec0c6e586bc59b2c2472601fbec8c2790551b9aa8e4ed77b4dbf0f4"}