{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:PQKYF5EROXD6TIUCZF73S5OBGL","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":"eb8721bd48a1c4ad4b948617da5c0d9a4ea11fee35ce0156ec52b8c6f890f7b4","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-12-15T18:49:43Z","title_canon_sha256":"35f81195bdf14cbe5bb53c572389843e339b06d5cf5f84f44081eed871e36edf"},"schema_version":"1.0","source":{"id":"2312.10029","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2312.10029","created_at":"2026-07-05T07:25:27Z"},{"alias_kind":"arxiv_version","alias_value":"2312.10029v2","created_at":"2026-07-05T07:25:27Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.10029","created_at":"2026-07-05T07:25:27Z"},{"alias_kind":"pith_short_12","alias_value":"PQKYF5EROXD6","created_at":"2026-07-05T07:25:27Z"},{"alias_kind":"pith_short_16","alias_value":"PQKYF5EROXD6TIUC","created_at":"2026-07-05T07:25:27Z"},{"alias_kind":"pith_short_8","alias_value":"PQKYF5ER","created_at":"2026-07-05T07:25:27Z"}],"graph_snapshots":[{"event_id":"sha256:855999fb3992e9f0ce082b7c16642415e39c9b4d1e07763dc90da426882c7e4e","target":"graph","created_at":"2026-07-05T07:25:27Z","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/2312.10029/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We show that existing unsupervised methods on large language model (LLM) activations do not discover knowledge -- instead they seem to discover whatever feature of the activations is most prominent. The idea behind unsupervised knowledge elicitation is that knowledge satisfies a consistency structure, which can be used to discover knowledge. We first prove theoretically that arbitrary features (not just knowledge) satisfy the consistency structure of a particular leading unsupervised knowledge-elicitation method, contrast-consistent search (Burns et al. - arXiv:2212.03827). We then present a s","authors_text":"Johannes Gasteiger, Rohin Shah, Sebastian Farquhar, Vikrant Varma, Vladimir Mikulik, Zachary Kenton","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-12-15T18:49:43Z","title":"Challenges with unsupervised LLM knowledge discovery"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.10029","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":"sha256:9d092a8e2e007efaf726e658aed8934ed3d9c9f41b9ec003ab9d89ef938206a4","target":"record","created_at":"2026-07-05T07:25:27Z","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":"eb8721bd48a1c4ad4b948617da5c0d9a4ea11fee35ce0156ec52b8c6f890f7b4","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-12-15T18:49:43Z","title_canon_sha256":"35f81195bdf14cbe5bb53c572389843e339b06d5cf5f84f44081eed871e36edf"},"schema_version":"1.0","source":{"id":"2312.10029","kind":"arxiv","version":2}},"canonical_sha256":"7c1582f49175c7e9a282c97fb975c132fb0168fed8fe2fcab10d74b94d38a1f2","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7c1582f49175c7e9a282c97fb975c132fb0168fed8fe2fcab10d74b94d38a1f2","first_computed_at":"2026-07-05T07:25:27.201580Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:25:27.201580Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"4x67VydWjOCUx9hnIZMXlWUeg+VapjGwB+AzykDqd43xIf4ojZw1pAEKrHKpmiP11vm0CCU4capAKxSrfz0vCA==","signature_status":"signed_v1","signed_at":"2026-07-05T07:25:27.202050Z","signed_message":"canonical_sha256_bytes"},"source_id":"2312.10029","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9d092a8e2e007efaf726e658aed8934ed3d9c9f41b9ec003ab9d89ef938206a4","sha256:855999fb3992e9f0ce082b7c16642415e39c9b4d1e07763dc90da426882c7e4e"],"state_sha256":"daeb99482754186d226c4f42d038b64cdcfa1296cb43b6d3d55f7f36643b2557"}