{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:H3N5WVVLRCXM54YPXSYWGQIYYJ","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":"97bd0de254b9c1cb3e854c2f26525ff26bc1cb401105aaaf7ec44ec942776dcd","cross_cats_sorted":["cs.AI","cs.CL","cs.ET","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2026-06-28T15:05:12Z","title_canon_sha256":"edf6456f8750aca30c87219139e99dd69e7eb269aa27ddffe354aca82801fb07"},"schema_version":"1.0","source":{"id":"2606.29441","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.29441","created_at":"2026-06-30T01:18:06Z"},{"alias_kind":"arxiv_version","alias_value":"2606.29441v1","created_at":"2026-06-30T01:18:06Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.29441","created_at":"2026-06-30T01:18:06Z"},{"alias_kind":"pith_short_12","alias_value":"H3N5WVVLRCXM","created_at":"2026-06-30T01:18:06Z"},{"alias_kind":"pith_short_16","alias_value":"H3N5WVVLRCXM54YP","created_at":"2026-06-30T01:18:06Z"},{"alias_kind":"pith_short_8","alias_value":"H3N5WVVL","created_at":"2026-06-30T01:18:06Z"}],"graph_snapshots":[{"event_id":"sha256:423c9f5c38818662dfbe45edd31803436782e2da6c3c7c4c8e6dd57c05c3eb8a","target":"graph","created_at":"2026-06-30T01:18:06Z","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/2606.29441/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Inference-time safety methods for large language models have proliferated, yet no systematic comparison exists. We evaluate five defense paradigms (no defense, static steering, CAST, AlphaSteer, probe-gated) across seven instruction-tuned models (7-31B) and five attack types (GCG, AutoDAN, DeepInception, prefilling, intent laundering). Our central finding: prompt-time activation defenses are structurally blind to prefilling attacks. AlphaSteer achieves 0% attack success on GCG, AutoDAN, and intent laundering but 50% on prefilling. We prove a corollary: any defense that gates intervention on a ","authors_text":"Subhadip Mitra","cross_cats":["cs.AI","cs.CL","cs.ET","cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2026-06-28T15:05:12Z","title":"Closing the Activation-Cone Blind Spot: Response-Time Probing and Unified Defense"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.29441","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":"sha256:42dc1db379f97c1cd42d4a5027db45c78b909caa6abf5815df800cd9308d9639","target":"record","created_at":"2026-06-30T01:18:06Z","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":"97bd0de254b9c1cb3e854c2f26525ff26bc1cb401105aaaf7ec44ec942776dcd","cross_cats_sorted":["cs.AI","cs.CL","cs.ET","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2026-06-28T15:05:12Z","title_canon_sha256":"edf6456f8750aca30c87219139e99dd69e7eb269aa27ddffe354aca82801fb07"},"schema_version":"1.0","source":{"id":"2606.29441","kind":"arxiv","version":1}},"canonical_sha256":"3edbdb56ab88aecef30fbcb1634118c250607b5dfdc90f20e1e152a10f8521f8","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3edbdb56ab88aecef30fbcb1634118c250607b5dfdc90f20e1e152a10f8521f8","first_computed_at":"2026-06-30T01:18:06.571405Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-30T01:18:06.571405Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"SDOJdv0io/zvFyOSc5N1vBHotRUDwx+2wo2aONPPr5FV0m1l5onmmlO4x3RdMWCVoav5uh0Yj0PNENtXd3LsBA==","signature_status":"signed_v1","signed_at":"2026-06-30T01:18:06.572084Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.29441","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:42dc1db379f97c1cd42d4a5027db45c78b909caa6abf5815df800cd9308d9639","sha256:423c9f5c38818662dfbe45edd31803436782e2da6c3c7c4c8e6dd57c05c3eb8a"],"state_sha256":"a989d75a87c0b2ce60ad1d2dfd58ab428ae15bc89531ee7332519ecbe155c761"}