{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:XZY4ZXL3XZ2D2LICCL4JKYQFZE","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":"cf1283eaae0d3c3f8260207ad798c8cb1550b8dc42bd189aacf550e82cc713e6","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-05-02T14:56:37Z","title_canon_sha256":"7a717b3cf54a281cb8222adc6d592e0a68a6c5804b8f43693ec264675e837ec2"},"schema_version":"1.0","source":{"id":"2605.02958","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.02958","created_at":"2026-05-27T02:05:20Z"},{"alias_kind":"arxiv_version","alias_value":"2605.02958v2","created_at":"2026-05-27T02:05:20Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.02958","created_at":"2026-05-27T02:05:20Z"},{"alias_kind":"pith_short_12","alias_value":"XZY4ZXL3XZ2D","created_at":"2026-05-27T02:05:20Z"},{"alias_kind":"pith_short_16","alias_value":"XZY4ZXL3XZ2D2LIC","created_at":"2026-05-27T02:05:20Z"},{"alias_kind":"pith_short_8","alias_value":"XZY4ZXL3","created_at":"2026-05-27T02:05:20Z"}],"graph_snapshots":[{"event_id":"sha256:c5e6785f7b627b9e2d91f647ebd8930d6e694c789b0c5ced42cfd148af5fe796","target":"graph","created_at":"2026-05-27T02:05:20Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"SALO effectively recovers defense capabilities against forced-decoding attacks, improving detection rates from ~0% to >90% where methods relying on terminal states perform poorly."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the refusal trajectory uncovered by causal tracing is a stable, generalizable signature rather than an artifact of the specific models, prompts, or attack implementations tested."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Refusal in LLMs leaves a detectable upstream trajectory that SALO exploits to raise jailbreak detection from near zero to over 90 percent even under forced-decoding attacks."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Refusal in language models follows a persistent upstream trajectory that remains detectable even when attacks suppress the final output."}],"snapshot_sha256":"98f8241d2d7d8505d4472f6174cf11b249fde10448a9087056e960d42b906000"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-20T17:41:08.913207Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T17:16:24.573853Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2605.02958/integrity.json","findings":[],"snapshot_sha256":"75ce4b50a859a889d1999293782c18672c4f12c631159e8e7cd1f0e05517dd4d","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Representation Engineering analyses often characterize refusal using static directions extracted from terminal or pooled representations. We ask whether this view misses how refusal is constructed across layer-token positions. Using causal tracing, we identify a \\textit{Refusal Trajectory}: a sparse upstream activation pattern that often persists even when attacks such as GCG suppress terminal refusal signals. Based on this observation, we propose SALO (Sparse Activation Localization Operator), a lightweight white-box detector that operates on raw hidden-state volumes from a selected layer win","authors_text":"Che Wang, Jianbo Gao, Wei Yang Bryan Lim, Xulin Hu, Zhong Chen","cross_cats":["cs.AI","cs.CL","cs.LG"],"headline":"Refusal in language models follows a persistent upstream trajectory that remains detectable even when attacks suppress the final output.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-05-02T14:56:37Z","title":"Tracing the Dynamics of Refusal: Exploiting Latent Refusal Trajectories for Robust Jailbreak Detection"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.02958","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-09T14:08:46.459279Z","id":"6cc86458-824d-4c8e-a196-1b6d49eba5b7","model_set":{"reader":"grok-4.3"},"one_line_summary":"Refusal in LLMs leaves a detectable upstream trajectory that SALO exploits to raise jailbreak detection from near zero to over 90 percent even under forced-decoding attacks.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Refusal in language models follows a persistent upstream trajectory that remains detectable even when attacks suppress the final output.","strongest_claim":"SALO effectively recovers defense capabilities against forced-decoding attacks, improving detection rates from ~0% to >90% where methods relying on terminal states perform poorly.","weakest_assumption":"That the refusal trajectory uncovered by causal tracing is a stable, generalizable signature rather than an artifact of the specific models, prompts, or attack implementations tested."}},"verdict_id":"6cc86458-824d-4c8e-a196-1b6d49eba5b7"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:0701645b75352bf65da8f8cc5caff8a9b7045cad31eb8035c6dc78e620522794","target":"record","created_at":"2026-05-27T02:05:20Z","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":"cf1283eaae0d3c3f8260207ad798c8cb1550b8dc42bd189aacf550e82cc713e6","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-05-02T14:56:37Z","title_canon_sha256":"7a717b3cf54a281cb8222adc6d592e0a68a6c5804b8f43693ec264675e837ec2"},"schema_version":"1.0","source":{"id":"2605.02958","kind":"arxiv","version":2}},"canonical_sha256":"be71ccdd7bbe743d2d0212f8956205c917e64bd2b553330318ae754b2634ae0a","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"be71ccdd7bbe743d2d0212f8956205c917e64bd2b553330318ae754b2634ae0a","first_computed_at":"2026-05-27T02:05:20.995842Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-27T02:05:20.995842Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"VHr4mqMrWRtokhaZiSpkB1gIf1/OrdvRAK7O2uu0L6Vp3efD9LxZPV5nmWgZ9QO3iHItgWZ69dTOoAeMO7g3Bw==","signature_status":"signed_v1","signed_at":"2026-05-27T02:05:20.996373Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.02958","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0701645b75352bf65da8f8cc5caff8a9b7045cad31eb8035c6dc78e620522794","sha256:c5e6785f7b627b9e2d91f647ebd8930d6e694c789b0c5ced42cfd148af5fe796"],"state_sha256":"271cf4adf4e2179e1f2478187a73c7931ec1308206669ed97cc1c1bbad30e175"}