{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:3BXSRL7NP6V4TC4VTAUQXHFTV5","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":"e32ee27bbd1e6b25ca37c9041433ead9bc3da6b0b2bdfdf7dbcdc69000188d62","cross_cats_sorted":["cs.CL","cs.CR","cs.SE"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-07-23T09:14:27Z","title_canon_sha256":"1ef6b7af1e93bad060df4c557f27d8532794008c251444f1a72caef10f05c172"},"schema_version":"1.0","source":{"id":"2407.16318","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2407.16318","created_at":"2026-07-05T08:47:31Z"},{"alias_kind":"arxiv_version","alias_value":"2407.16318v1","created_at":"2026-07-05T08:47:31Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.16318","created_at":"2026-07-05T08:47:31Z"},{"alias_kind":"pith_short_12","alias_value":"3BXSRL7NP6V4","created_at":"2026-07-05T08:47:31Z"},{"alias_kind":"pith_short_16","alias_value":"3BXSRL7NP6V4TC4V","created_at":"2026-07-05T08:47:31Z"},{"alias_kind":"pith_short_8","alias_value":"3BXSRL7N","created_at":"2026-07-05T08:47:31Z"}],"graph_snapshots":[{"event_id":"sha256:68e1f60bf36932c28e16e7259e894e16fc9f36bc2a159507a97abbc6a4a214b4","target":"graph","created_at":"2026-07-05T08:47:31Z","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/2407.16318/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Deploying language models (LMs) necessitates outputs to be both high-quality and compliant with safety guidelines. Although Inference-Time Guardrails (ITG) offer solutions that shift model output distributions towards compliance, we find that current methods struggle in balancing safety with helpfulness. ITG Methods that safely address non-compliant queries exhibit lower helpfulness while those that prioritize helpfulness compromise on safety. We refer to this trade-off as the guardrail tax, analogous to the alignment tax. To address this, we propose PrimeGuard, a novel ITG method that utilize","authors_text":"Blazej Manczak, Eliott Zemour, Eric Lin, Vaikkunth Mugunthan","cross_cats":["cs.CL","cs.CR","cs.SE"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-07-23T09:14:27Z","title":"PrimeGuard: Safe and Helpful LLMs through Tuning-Free Routing"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.16318","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":"sha256:4d071b82d7d16ae14d252a87f3d9c59834a27985dbbf75585ac0cf016fe488c6","target":"record","created_at":"2026-07-05T08:47:31Z","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":"e32ee27bbd1e6b25ca37c9041433ead9bc3da6b0b2bdfdf7dbcdc69000188d62","cross_cats_sorted":["cs.CL","cs.CR","cs.SE"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-07-23T09:14:27Z","title_canon_sha256":"1ef6b7af1e93bad060df4c557f27d8532794008c251444f1a72caef10f05c172"},"schema_version":"1.0","source":{"id":"2407.16318","kind":"arxiv","version":1}},"canonical_sha256":"d86f28afed7fabc98b9598290b9cb3af4805a39973c7424243332e7346d31a9f","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"d86f28afed7fabc98b9598290b9cb3af4805a39973c7424243332e7346d31a9f","first_computed_at":"2026-07-05T08:47:31.363639Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:47:31.363639Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"13kTmvyZp4b9u++Cxj/BiyuFnQtW9tSKckXX7q4Q9yY+yLSlVjgyj+I6lma/7T+VX+jv15k1fv6haFrVZ5xBAA==","signature_status":"signed_v1","signed_at":"2026-07-05T08:47:31.364124Z","signed_message":"canonical_sha256_bytes"},"source_id":"2407.16318","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4d071b82d7d16ae14d252a87f3d9c59834a27985dbbf75585ac0cf016fe488c6","sha256:68e1f60bf36932c28e16e7259e894e16fc9f36bc2a159507a97abbc6a4a214b4"],"state_sha256":"672142f957bbc84832ac0c43cae73b7830f6a85037b0a028f8d4f1531c0d283d"}