{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:GS2YOPBIMHCOIDQJ4BC4G4TUCN","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":"6313eecb8297f767366d258a09c5fe29a8037beb822399826c23f8fe22cee35e","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-02-23T18:19:46Z","title_canon_sha256":"17024bd19fcc5327c5f7178f07e03c7c5209ad9ca46ae0ee04ce26b49aff5bd2"},"schema_version":"1.0","source":{"id":"2602.20102","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.20102","created_at":"2026-05-25T02:01:15Z"},{"alias_kind":"arxiv_version","alias_value":"2602.20102v2","created_at":"2026-05-25T02:01:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.20102","created_at":"2026-05-25T02:01:15Z"},{"alias_kind":"pith_short_12","alias_value":"GS2YOPBIMHCO","created_at":"2026-05-25T02:01:15Z"},{"alias_kind":"pith_short_16","alias_value":"GS2YOPBIMHCOIDQJ","created_at":"2026-05-25T02:01:15Z"},{"alias_kind":"pith_short_8","alias_value":"GS2YOPBI","created_at":"2026-05-25T02:01:15Z"}],"graph_snapshots":[{"event_id":"sha256:467d7e5966521a0dc91c5610544c611cdb930850fd3befc1e8f370594ad51465","target":"graph","created_at":"2026-05-25T02:01:15Z","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/2602.20102/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Despite the strong performance of large language models (LLMs) across diverse tasks, their susceptibility to adversarial attacks and unsafe content generation remains a significant obstacle to deployment, particularly in high-stakes settings. Addressing this challenge requires safety mechanisms that are both practically effective and theoretically grounded. In this paper, we introduce BarrierSteer, a novel inference-time framework that improves response safety by embedding learned nonlinear safety constraints directly into the model's latent representation space. BarrierSteer treats hidden-sta","authors_text":"Arun Verma, Bryan Kian Hsiang Low, Daniela Rus, Kiwan Wong, Thanh Q. Tran, Wei Xiao","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-02-23T18:19:46Z","title":"BarrierSteer: LLM Safety via Learning Barrier Steering"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.20102","kind":"arxiv","version":2},"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:bc2d96bb0ce82d98d29916940ca3b1745aa55cc0a3d9a31ce085d0f7159d5a62","target":"record","created_at":"2026-05-25T02:01:15Z","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":"6313eecb8297f767366d258a09c5fe29a8037beb822399826c23f8fe22cee35e","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-02-23T18:19:46Z","title_canon_sha256":"17024bd19fcc5327c5f7178f07e03c7c5209ad9ca46ae0ee04ce26b49aff5bd2"},"schema_version":"1.0","source":{"id":"2602.20102","kind":"arxiv","version":2}},"canonical_sha256":"34b5873c2861c4e40e09e045c37274136d4e84c58e7a7c1dfba3479895511995","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"34b5873c2861c4e40e09e045c37274136d4e84c58e7a7c1dfba3479895511995","first_computed_at":"2026-05-25T02:01:15.959847Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-25T02:01:15.959847Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"szicRJS7Ol4guNiLCBQxkoAAxChA9xCbRFEksX3G2Asvh4RyC3gXbKmhdqParMXfMfcxhNCGaOoPwChE2q+CAA==","signature_status":"signed_v1","signed_at":"2026-05-25T02:01:15.960581Z","signed_message":"canonical_sha256_bytes"},"source_id":"2602.20102","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:bc2d96bb0ce82d98d29916940ca3b1745aa55cc0a3d9a31ce085d0f7159d5a62","sha256:467d7e5966521a0dc91c5610544c611cdb930850fd3befc1e8f370594ad51465"],"state_sha256":"76e6af14c60bbd2c4dcd341014da1b95ba5f517f01a07009a14d959c6629d757"}