{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:3DAI3M6PTNBQY6FQJUWNEV6MWI","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":"e8871745350361d057329eda5a4f05e34e8ec93e3cfba4aafe0a9f07b00eea22","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-27T13:33:26Z","title_canon_sha256":"a47abab0963673886cf60bb7ce5e6059c18dbc77e47eebc6110d6ab32a389497"},"schema_version":"1.0","source":{"id":"2605.28467","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.28467","created_at":"2026-05-28T02:04:53Z"},{"alias_kind":"arxiv_version","alias_value":"2605.28467v1","created_at":"2026-05-28T02:04:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.28467","created_at":"2026-05-28T02:04:53Z"},{"alias_kind":"pith_short_12","alias_value":"3DAI3M6PTNBQ","created_at":"2026-05-28T02:04:53Z"},{"alias_kind":"pith_short_16","alias_value":"3DAI3M6PTNBQY6FQ","created_at":"2026-05-28T02:04:53Z"},{"alias_kind":"pith_short_8","alias_value":"3DAI3M6P","created_at":"2026-05-28T02:04:53Z"}],"graph_snapshots":[{"event_id":"sha256:982355a22a0a672dbc4ee920dda0f60f9fbdf1794d833786cd5b76753445f025","target":"graph","created_at":"2026-05-28T02:04:53Z","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/2605.28467/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"As LLMs gain stronger reasoning capabilities, their extended chain-of-thought introduces new degrees of complexity for defending against adversarial jailbreaks and prompt injection. We study consistency training, a family of fine-tuning objectives that enforce identical behavior on clean prompts and adversarial rewrites, and evaluate its two main variants, output-level (BCT) and activation-level (ACT), across five reasoning models. We formulate both methods as a prompt injection defense and find ACT to be competitive with other training-based defenses while requiring only self-supervised pairs","authors_text":"Avidan Shah, Jannik Brinkmann, Rico Angell","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-27T13:33:26Z","title":"Mitigating Adaptive Attacks against Reasoning Models with Activation Consistency Training"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.28467","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:fb12c0377f2a1c809aa541c2acc3024cca9551111b9d71322fafba65e56e6549","target":"record","created_at":"2026-05-28T02:04:53Z","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":"e8871745350361d057329eda5a4f05e34e8ec93e3cfba4aafe0a9f07b00eea22","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-27T13:33:26Z","title_canon_sha256":"a47abab0963673886cf60bb7ce5e6059c18dbc77e47eebc6110d6ab32a389497"},"schema_version":"1.0","source":{"id":"2605.28467","kind":"arxiv","version":1}},"canonical_sha256":"d8c08db3cf9b430c78b04d2cd257ccb22024449f130e7593e1abac08e3823214","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"d8c08db3cf9b430c78b04d2cd257ccb22024449f130e7593e1abac08e3823214","first_computed_at":"2026-05-28T02:04:53.880429Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-28T02:04:53.880429Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"dtR8ScuGh2JBEShRHvGAvWwfQrI12FaVckxfv3LsL8budlchJ+iiyKS41rdutwKCQq48ZJ6VNZ+uxhc9thPQAA==","signature_status":"signed_v1","signed_at":"2026-05-28T02:04:53.880815Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.28467","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:fb12c0377f2a1c809aa541c2acc3024cca9551111b9d71322fafba65e56e6549","sha256:982355a22a0a672dbc4ee920dda0f60f9fbdf1794d833786cd5b76753445f025"],"state_sha256":"e76b7d8e6c529078032bc1e1970854a8603dd1db83f0fa5f168b8a8c540511a5"}