{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:NOZTD4BKYNODWZSPMXU4KVNXAJ","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":"ae037134bb805ed5dd727df3ab3ff1d678504848b20f84331178c44618954847","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-18T01:14:46Z","title_canon_sha256":"0f4aab8fad5e6aa381f0cc00d6604c048a01801a24c6686c3438ee29eef492ef"},"schema_version":"1.0","source":{"id":"2502.12411","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.12411","created_at":"2026-07-05T10:15:55Z"},{"alias_kind":"arxiv_version","alias_value":"2502.12411v1","created_at":"2026-07-05T10:15:55Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.12411","created_at":"2026-07-05T10:15:55Z"},{"alias_kind":"pith_short_12","alias_value":"NOZTD4BKYNOD","created_at":"2026-07-05T10:15:55Z"},{"alias_kind":"pith_short_16","alias_value":"NOZTD4BKYNODWZSP","created_at":"2026-07-05T10:15:55Z"},{"alias_kind":"pith_short_8","alias_value":"NOZTD4BK","created_at":"2026-07-05T10:15:55Z"}],"graph_snapshots":[{"event_id":"sha256:20657331f9003054d6ac3beab06083b038f2617fc6fb6af0c1aeac8610b6fc41","target":"graph","created_at":"2026-07-05T10:15:55Z","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/2502.12411/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Unsafe prompts pose significant safety risks to large language models (LLMs). Existing methods for detecting unsafe prompts rely on data-driven fine-tuning to train guardrail models, necessitating significant data and computational resources. In contrast, recent few-shot gradient-based methods emerge, requiring only few safe and unsafe reference prompts. A gradient-based approach identifies unsafe prompts by analyzing consistent patterns of the gradients of safety-critical parameters in LLMs. Although effective, its restriction to directional similarity (cosine similarity) introduces ``directi","authors_text":"Bowen Yan, Jingyuan Yang, Rongjun Li, Wei Peng, Xin Chen, Zhiyong Feng, Ziyu Zhou","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-18T01:14:46Z","title":"Gradient Co-occurrence Analysis for Detecting Unsafe Prompts in Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.12411","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":"sha256:58cd7a5628718a734f7feba4df06d4cb427b76f128d7ecdf071bbfdf152d2a56","target":"record","created_at":"2026-07-05T10:15:55Z","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":"ae037134bb805ed5dd727df3ab3ff1d678504848b20f84331178c44618954847","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-18T01:14:46Z","title_canon_sha256":"0f4aab8fad5e6aa381f0cc00d6604c048a01801a24c6686c3438ee29eef492ef"},"schema_version":"1.0","source":{"id":"2502.12411","kind":"arxiv","version":1}},"canonical_sha256":"6bb331f02ac35c3b664f65e9c555b7024c663a242d53feaa2e173ebc747f6276","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"6bb331f02ac35c3b664f65e9c555b7024c663a242d53feaa2e173ebc747f6276","first_computed_at":"2026-07-05T10:15:55.413818Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:15:55.413818Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"/ODaYq1AshokpQDrivbfupmPhjLlKO1ZWM2DrAA560LdVtveuKlfgWmGIyblACkvechqYNwZpJ2+2p+iUaAZDg==","signature_status":"signed_v1","signed_at":"2026-07-05T10:15:55.414271Z","signed_message":"canonical_sha256_bytes"},"source_id":"2502.12411","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:58cd7a5628718a734f7feba4df06d4cb427b76f128d7ecdf071bbfdf152d2a56","sha256:20657331f9003054d6ac3beab06083b038f2617fc6fb6af0c1aeac8610b6fc41"],"state_sha256":"e259701a96de4719a249172ca16c1d30f499c0b56749dea3007b0e35aa16a22f"}