{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:GFIFYHZZQLCX67PTXEO7TYD35G","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":"031e02220c1b3b424ad7affcbe95747d2aa9e19bf8c6ce964009ae506a68ed95","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-02-01T20:18:11Z","title_canon_sha256":"6d3b87413fa46fdfaa842ffcf73d08a38d79ff4247ab10e464beac1757024557"},"schema_version":"1.0","source":{"id":"2602.01425","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.01425","created_at":"2026-06-19T16:12:18Z"},{"alias_kind":"arxiv_version","alias_value":"2602.01425v2","created_at":"2026-06-19T16:12:18Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.01425","created_at":"2026-06-19T16:12:18Z"},{"alias_kind":"pith_short_12","alias_value":"GFIFYHZZQLCX","created_at":"2026-06-19T16:12:18Z"},{"alias_kind":"pith_short_16","alias_value":"GFIFYHZZQLCX67PT","created_at":"2026-06-19T16:12:18Z"},{"alias_kind":"pith_short_8","alias_value":"GFIFYHZZ","created_at":"2026-06-19T16:12:18Z"}],"graph_snapshots":[{"event_id":"sha256:72a39587f6c296d0b801c12cd87379060d71a44ab59f8e585cb8cdb5fe337208","target":"graph","created_at":"2026-06-19T16:12:18Z","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/2602.01425/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Linear probes are a promising approach for monitoring AI systems for deceptive behaviour. Previous work has shown that a linear classifier trained on a contrastive instruction pair and a simple dataset can achieve good performance. However, these probes exhibit notable failures even in straightforward scenarios, including spurious correlations and false positives on non-deceptive responses. In this paper, we demonstrate that deception detection is inherently heterogeneous: while a single universal probe achieves modest improvements (+0.032 AUC), post-hoc oracle analysis reveals substantially h","authors_text":"Devina Jain, Joseph Bloom, Satvik Golechha, Shivam Arora, Vikram Natarajan","cross_cats":["cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-02-01T20:18:11Z","title":"One Probe Won't Catch Them All: Towards Targeted Deception Detection"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.01425","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:bfdb48e10ee6a716a47dadea035edcaa167aa905504d55fab39a237616e4bb1a","target":"record","created_at":"2026-06-19T16:12:18Z","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":"031e02220c1b3b424ad7affcbe95747d2aa9e19bf8c6ce964009ae506a68ed95","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-02-01T20:18:11Z","title_canon_sha256":"6d3b87413fa46fdfaa842ffcf73d08a38d79ff4247ab10e464beac1757024557"},"schema_version":"1.0","source":{"id":"2602.01425","kind":"arxiv","version":2}},"canonical_sha256":"31505c1f3982c57f7df3b91df9e07be9b693fb3f94480e7e556651e8b2d6b847","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"31505c1f3982c57f7df3b91df9e07be9b693fb3f94480e7e556651e8b2d6b847","first_computed_at":"2026-06-19T16:12:18.258466Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-19T16:12:18.258466Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"rznwDMPHzZF7/EawXndWTXhn/5iirdL5WqR5e0jHL1DNf8vcmDNzU/sTIIWpirxGckm3yvO3MvfNP4l9LWPOAQ==","signature_status":"signed_v1","signed_at":"2026-06-19T16:12:18.258882Z","signed_message":"canonical_sha256_bytes"},"source_id":"2602.01425","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:bfdb48e10ee6a716a47dadea035edcaa167aa905504d55fab39a237616e4bb1a","sha256:72a39587f6c296d0b801c12cd87379060d71a44ab59f8e585cb8cdb5fe337208"],"state_sha256":"a694586ca01c5236680f66df8eb624ca2c4d8c079bf655c5dc45136696fc199b"}