{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:S2EXWZWRVNQRZBLRPZYMYRDX54","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":"ea16bb990772f2095d0a91e82d468b21ad42d68b3a62e378ec250ac531337849","cross_cats_sorted":["cs.AI","cs.CR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-30T07:39:42Z","title_canon_sha256":"0a95a57701417c57eeb5c2b988b6d772c72ebb93ca89f5713e6f557cb2af5af6"},"schema_version":"1.0","source":{"id":"2410.22770","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.22770","created_at":"2026-07-05T10:41:16Z"},{"alias_kind":"arxiv_version","alias_value":"2410.22770v3","created_at":"2026-07-05T10:41:16Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.22770","created_at":"2026-07-05T10:41:16Z"},{"alias_kind":"pith_short_12","alias_value":"S2EXWZWRVNQR","created_at":"2026-07-05T10:41:16Z"},{"alias_kind":"pith_short_16","alias_value":"S2EXWZWRVNQRZBLR","created_at":"2026-07-05T10:41:16Z"},{"alias_kind":"pith_short_8","alias_value":"S2EXWZWR","created_at":"2026-07-05T10:41:16Z"}],"graph_snapshots":[{"event_id":"sha256:3c434887a4e7ff8022c23b45bb7736a0ff694a240ec2bcbf5b54c208659fe671","target":"graph","created_at":"2026-07-05T10:41:16Z","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/2410.22770/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Prompt injection attacks pose a critical threat to large language models (LLMs), enabling goal hijacking and data leakage. Prompt guard models, though effective in defense, suffer from over-defense -- falsely flagging benign inputs as malicious due to trigger word bias. To address this issue, we introduce NotInject, an evaluation dataset that systematically measures over-defense across various prompt guard models. NotInject contains 339 benign samples enriched with trigger words common in prompt injection attacks, enabling fine-grained evaluation. Our results show that state-of-the-art models ","authors_text":"Hao Li, Xiaogeng Liu","cross_cats":["cs.AI","cs.CR"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-30T07:39:42Z","title":"InjecGuard: Benchmarking and Mitigating Over-defense in Prompt Injection Guardrail Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.22770","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":"sha256:06936754e635bb64d28c3104b963dcccfc5278bee8dce17b011dcc2c378e7142","target":"record","created_at":"2026-07-05T10:41:16Z","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":"ea16bb990772f2095d0a91e82d468b21ad42d68b3a62e378ec250ac531337849","cross_cats_sorted":["cs.AI","cs.CR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-30T07:39:42Z","title_canon_sha256":"0a95a57701417c57eeb5c2b988b6d772c72ebb93ca89f5713e6f557cb2af5af6"},"schema_version":"1.0","source":{"id":"2410.22770","kind":"arxiv","version":3}},"canonical_sha256":"96897b66d1ab611c85717e70cc4477ef2f2e563dcdfcda6e14e27a992c89c326","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"96897b66d1ab611c85717e70cc4477ef2f2e563dcdfcda6e14e27a992c89c326","first_computed_at":"2026-07-05T10:41:16.766321Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:41:16.766321Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"0dcbCnBhrZ3gPLJkOAt4vMUKhIXNQBPUiN/YQrSnilBNizkE/ekkxbIMbG3YZwN0uqqXuZ2S4ejdQJrQ7/PmCw==","signature_status":"signed_v1","signed_at":"2026-07-05T10:41:16.766904Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.22770","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:06936754e635bb64d28c3104b963dcccfc5278bee8dce17b011dcc2c378e7142","sha256:3c434887a4e7ff8022c23b45bb7736a0ff694a240ec2bcbf5b54c208659fe671"],"state_sha256":"a947999b75bc94b341cb6840beb6e24319e0d4ec0a6602ad1162cbb2066bb10b"}