{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:7FNIQNELJW2QEFLE3AHNDONKBG","short_pith_number":"pith:7FNIQNEL","schema_version":"1.0","canonical_sha256":"f95a88348b4db5021564d80ed1b9aa09a3fdbcda6e91c7a86e1882848d19ffe7","source":{"kind":"arxiv","id":"2605.27958","version":1},"attestation_state":"computed","paper":{"title":"Pressure-Testing Deception Probes in LLMs: Scaling, Robustness, and the Geometry of Deceptive Representations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Sachin Kumar","submitted_at":"2026-05-27T04:51:55Z","abstract_excerpt":"Linear probes trained on LLM activations are increasingly proposed as deception-detection metrics, yet report AUROC exceeding 0.96 on clean benchmarks while collapsing under distributional shift. This paper systematically pressure-tests probe-based metrics across the Gemma 3 model family (1B-27B parameters), diagnosing why they fail rather than merely documenting that they fail. We test four hypotheses about deception encoding: (1) single linear direction, (2) multi-dimensional subspace, (3) convex conic hull, (4) entropy proxy. Our design includes cross-domain transfer matrices, multi-dimensi"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2605.27958","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-27T04:51:55Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"aa0051dbe191a5862e5089ef95e9880b4930694eead61469cbcc057bec724e10","abstract_canon_sha256":"00ad2851a0e1a580d4659ebf2fe45ba15b4060ee274c8048ad9c06e3f9cad948"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:04:54.160710Z","signature_b64":"dqQfuv/eKqtrBpzNxQKPtIwJoWwBS0C5KroCvpCygNcPl9LqxDj7letbGTR2W9+GGc0/w/Cr+yH9vp3YNrtfBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f95a88348b4db5021564d80ed1b9aa09a3fdbcda6e91c7a86e1882848d19ffe7","last_reissued_at":"2026-05-28T01:04:54.160289Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:04:54.160289Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Pressure-Testing Deception Probes in LLMs: Scaling, Robustness, and the Geometry of Deceptive Representations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Sachin Kumar","submitted_at":"2026-05-27T04:51:55Z","abstract_excerpt":"Linear probes trained on LLM activations are increasingly proposed as deception-detection metrics, yet report AUROC exceeding 0.96 on clean benchmarks while collapsing under distributional shift. This paper systematically pressure-tests probe-based metrics across the Gemma 3 model family (1B-27B parameters), diagnosing why they fail rather than merely documenting that they fail. We test four hypotheses about deception encoding: (1) single linear direction, (2) multi-dimensional subspace, (3) convex conic hull, (4) entropy proxy. Our design includes cross-domain transfer matrices, multi-dimensi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.27958","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.27958/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2605.27958","created_at":"2026-05-28T01:04:54.160358+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.27958v1","created_at":"2026-05-28T01:04:54.160358+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.27958","created_at":"2026-05-28T01:04:54.160358+00:00"},{"alias_kind":"pith_short_12","alias_value":"7FNIQNELJW2Q","created_at":"2026-05-28T01:04:54.160358+00:00"},{"alias_kind":"pith_short_16","alias_value":"7FNIQNELJW2QEFLE","created_at":"2026-05-28T01:04:54.160358+00:00"},{"alias_kind":"pith_short_8","alias_value":"7FNIQNEL","created_at":"2026-05-28T01:04:54.160358+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7FNIQNELJW2QEFLE3AHNDONKBG","json":"https://pith.science/pith/7FNIQNELJW2QEFLE3AHNDONKBG.json","graph_json":"https://pith.science/api/pith-number/7FNIQNELJW2QEFLE3AHNDONKBG/graph.json","events_json":"https://pith.science/api/pith-number/7FNIQNELJW2QEFLE3AHNDONKBG/events.json","paper":"https://pith.science/paper/7FNIQNEL"},"agent_actions":{"view_html":"https://pith.science/pith/7FNIQNELJW2QEFLE3AHNDONKBG","download_json":"https://pith.science/pith/7FNIQNELJW2QEFLE3AHNDONKBG.json","view_paper":"https://pith.science/paper/7FNIQNEL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.27958&json=true","fetch_graph":"https://pith.science/api/pith-number/7FNIQNELJW2QEFLE3AHNDONKBG/graph.json","fetch_events":"https://pith.science/api/pith-number/7FNIQNELJW2QEFLE3AHNDONKBG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7FNIQNELJW2QEFLE3AHNDONKBG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7FNIQNELJW2QEFLE3AHNDONKBG/action/storage_attestation","attest_author":"https://pith.science/pith/7FNIQNELJW2QEFLE3AHNDONKBG/action/author_attestation","sign_citation":"https://pith.science/pith/7FNIQNELJW2QEFLE3AHNDONKBG/action/citation_signature","submit_replication":"https://pith.science/pith/7FNIQNELJW2QEFLE3AHNDONKBG/action/replication_record"}},"created_at":"2026-05-28T01:04:54.160358+00:00","updated_at":"2026-05-28T01:04:54.160358+00:00"}