{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:OWKSAWTK2FGMHON4UGZINRKQIV","short_pith_number":"pith:OWKSAWTK","schema_version":"1.0","canonical_sha256":"7595205a6ad14cc3b9bca1b286c550455dd9cfac6749ba127765833eb0079771","source":{"kind":"arxiv","id":"2607.02072","version":1},"attestation_state":"computed","paper":{"title":"kNNGuard: Turning LLM Hidden Activations into a Training-Free Configurable Guardrail","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CR"],"primary_cat":"cs.LG","authors_text":"Hamid Nasiri, Mahmoud Abdelfattah, Peter Garraghan","submitted_at":"2026-07-02T12:07:00Z","abstract_excerpt":"Large language models (LLMs) are increasingly deployed in domains requiring guardrails to detect unsafe, off-topic, or adversarial prompts. Existing guardrails predominately rely on fine-tuning to build classifiers, which often suffer from low generalization and high inference latency. We present kNNGuard, a training-free guardrail that utilizes the activation space of an off-the-shelf LLM. Given a small bank of 50 safe and unsafe prompts, kNNGuard extracts hidden activations and performs multi-layer kNN fusing activation-space and embedding-space scores for classification. Across six domains "},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2607.02072","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-02T12:07:00Z","cross_cats_sorted":["cs.AI","cs.CR"],"title_canon_sha256":"d43bd3329558fd7d0293419a59cc61dc3d92b13de6e200acab7bba8ce332aa40","abstract_canon_sha256":"ee7d9ba85132965606cf5e0837019582470ccdb4eb629a6e77c93c4865ec3e62"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-03T01:17:38.960793Z","signature_b64":"Xyd5zqrluo6Bq1VhbeXwU0kh3/LAXc0b2g8HVIo/4Chh+JH2UoOrqKoJcmY8v+59HxRSgkeWdQEJ8hpu5P/8BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7595205a6ad14cc3b9bca1b286c550455dd9cfac6749ba127765833eb0079771","last_reissued_at":"2026-07-03T01:17:38.960361Z","signature_status":"signed_v1","first_computed_at":"2026-07-03T01:17:38.960361Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"kNNGuard: Turning LLM Hidden Activations into a Training-Free Configurable Guardrail","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CR"],"primary_cat":"cs.LG","authors_text":"Hamid Nasiri, Mahmoud Abdelfattah, Peter Garraghan","submitted_at":"2026-07-02T12:07:00Z","abstract_excerpt":"Large language models (LLMs) are increasingly deployed in domains requiring guardrails to detect unsafe, off-topic, or adversarial prompts. Existing guardrails predominately rely on fine-tuning to build classifiers, which often suffer from low generalization and high inference latency. We present kNNGuard, a training-free guardrail that utilizes the activation space of an off-the-shelf LLM. Given a small bank of 50 safe and unsafe prompts, kNNGuard extracts hidden activations and performs multi-layer kNN fusing activation-space and embedding-space scores for classification. Across six domains "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.02072","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2607.02072/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2607.02072","created_at":"2026-07-03T01:17:38.960429+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.02072v1","created_at":"2026-07-03T01:17:38.960429+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.02072","created_at":"2026-07-03T01:17:38.960429+00:00"},{"alias_kind":"pith_short_12","alias_value":"OWKSAWTK2FGM","created_at":"2026-07-03T01:17:38.960429+00:00"},{"alias_kind":"pith_short_16","alias_value":"OWKSAWTK2FGMHON4","created_at":"2026-07-03T01:17:38.960429+00:00"},{"alias_kind":"pith_short_8","alias_value":"OWKSAWTK","created_at":"2026-07-03T01:17:38.960429+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OWKSAWTK2FGMHON4UGZINRKQIV","json":"https://pith.science/pith/OWKSAWTK2FGMHON4UGZINRKQIV.json","graph_json":"https://pith.science/api/pith-number/OWKSAWTK2FGMHON4UGZINRKQIV/graph.json","events_json":"https://pith.science/api/pith-number/OWKSAWTK2FGMHON4UGZINRKQIV/events.json","paper":"https://pith.science/paper/OWKSAWTK"},"agent_actions":{"view_html":"https://pith.science/pith/OWKSAWTK2FGMHON4UGZINRKQIV","download_json":"https://pith.science/pith/OWKSAWTK2FGMHON4UGZINRKQIV.json","view_paper":"https://pith.science/paper/OWKSAWTK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.02072&json=true","fetch_graph":"https://pith.science/api/pith-number/OWKSAWTK2FGMHON4UGZINRKQIV/graph.json","fetch_events":"https://pith.science/api/pith-number/OWKSAWTK2FGMHON4UGZINRKQIV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OWKSAWTK2FGMHON4UGZINRKQIV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OWKSAWTK2FGMHON4UGZINRKQIV/action/storage_attestation","attest_author":"https://pith.science/pith/OWKSAWTK2FGMHON4UGZINRKQIV/action/author_attestation","sign_citation":"https://pith.science/pith/OWKSAWTK2FGMHON4UGZINRKQIV/action/citation_signature","submit_replication":"https://pith.science/pith/OWKSAWTK2FGMHON4UGZINRKQIV/action/replication_record"}},"created_at":"2026-07-03T01:17:38.960429+00:00","updated_at":"2026-07-03T01:17:38.960429+00:00"}