{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:RJQ7OPHVBB7DVOD5VQNCLVJ2ZN","short_pith_number":"pith:RJQ7OPHV","canonical_record":{"source":{"id":"2405.01576","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-25T17:29:53Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"900d0815bbec065520f6caeff199fdda4568da7a9fcc96092243ba6a14d5d48e","abstract_canon_sha256":"9290bf8c52d3b0a5d2dbee9702b76c28576dd2f8ec1d8fc1e40e6f4bc694676c"},"schema_version":"1.0"},"canonical_sha256":"8a61f73cf5087e3ab87dac1a25d53acb56bc9c98f5ce618ce406e239a9f74817","source":{"kind":"arxiv","id":"2405.01576","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.01576","created_at":"2026-07-05T08:14:58Z"},{"alias_kind":"arxiv_version","alias_value":"2405.01576v1","created_at":"2026-07-05T08:14:58Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.01576","created_at":"2026-07-05T08:14:58Z"},{"alias_kind":"pith_short_12","alias_value":"RJQ7OPHVBB7D","created_at":"2026-07-05T08:14:58Z"},{"alias_kind":"pith_short_16","alias_value":"RJQ7OPHVBB7DVOD5","created_at":"2026-07-05T08:14:58Z"},{"alias_kind":"pith_short_8","alias_value":"RJQ7OPHV","created_at":"2026-07-05T08:14:58Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:RJQ7OPHVBB7DVOD5VQNCLVJ2ZN","target":"record","payload":{"canonical_record":{"source":{"id":"2405.01576","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-25T17:29:53Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"900d0815bbec065520f6caeff199fdda4568da7a9fcc96092243ba6a14d5d48e","abstract_canon_sha256":"9290bf8c52d3b0a5d2dbee9702b76c28576dd2f8ec1d8fc1e40e6f4bc694676c"},"schema_version":"1.0"},"canonical_sha256":"8a61f73cf5087e3ab87dac1a25d53acb56bc9c98f5ce618ce406e239a9f74817","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:14:58.257915Z","signature_b64":"QYLtqxPUdODZHqHHy+QarR1qkmNx1XezwEyeCACpjo8KKpst4+YocuPVv8Ntp9PEezRZqx7pbe6AqG2U6SnXAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8a61f73cf5087e3ab87dac1a25d53acb56bc9c98f5ce618ce406e239a9f74817","last_reissued_at":"2026-07-05T08:14:58.257486Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:14:58.257486Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.01576","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:14:58Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"SIYqiNtC/HXUiYhG//KSqU1eRrWCjhUfOD65XvhTL4BrUUpg5UczNG42fMgUbKro1hwveWYvhHdFbtWtA5V8CQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T02:27:48.806116Z"},"content_sha256":"7f863e4cf46d35f3d88f0778c89ca0a0fa44624e6de9393bc7b62985f231ff93","schema_version":"1.0","event_id":"sha256:7f863e4cf46d35f3d88f0778c89ca0a0fa44624e6de9393bc7b62985f231ff93"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:RJQ7OPHVBB7DVOD5VQNCLVJ2ZN","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Uncovering Deceptive Tendencies in Language Models: A Simulated Company AI Assistant","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Evan Hubinger, Olli J\\\"arviniemi","submitted_at":"2024-04-25T17:29:53Z","abstract_excerpt":"We study the tendency of AI systems to deceive by constructing a realistic simulation setting of a company AI assistant. The simulated company employees provide tasks for the assistant to complete, these tasks spanning writing assistance, information retrieval and programming. We then introduce situations where the model might be inclined to behave deceptively, while taking care to not instruct or otherwise pressure the model to do so. Across different scenarios, we find that Claude 3 Opus\n  1) complies with a task of mass-generating comments to influence public perception of the company, late"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.01576","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/2405.01576/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:14:58Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"dZy9nGbkeNoyddercdyQx426jsm4uQpk04rB/dMqcxp/QpymKco1LpZZ9m1AR+I4oxQfDD5VCUm4Q8i1iHqQCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T02:27:48.807030Z"},"content_sha256":"19f0d88701c1e2aa15497704a8650939cdfae279e8bbcd71a07c6443a6efb966","schema_version":"1.0","event_id":"sha256:19f0d88701c1e2aa15497704a8650939cdfae279e8bbcd71a07c6443a6efb966"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/RJQ7OPHVBB7DVOD5VQNCLVJ2ZN/bundle.json","state_url":"https://pith.science/pith/RJQ7OPHVBB7DVOD5VQNCLVJ2ZN/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/RJQ7OPHVBB7DVOD5VQNCLVJ2ZN/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-05T02:27:48Z","links":{"resolver":"https://pith.science/pith/RJQ7OPHVBB7DVOD5VQNCLVJ2ZN","bundle":"https://pith.science/pith/RJQ7OPHVBB7DVOD5VQNCLVJ2ZN/bundle.json","state":"https://pith.science/pith/RJQ7OPHVBB7DVOD5VQNCLVJ2ZN/state.json","well_known_bundle":"https://pith.science/.well-known/pith/RJQ7OPHVBB7DVOD5VQNCLVJ2ZN/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:RJQ7OPHVBB7DVOD5VQNCLVJ2ZN","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":"9290bf8c52d3b0a5d2dbee9702b76c28576dd2f8ec1d8fc1e40e6f4bc694676c","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-25T17:29:53Z","title_canon_sha256":"900d0815bbec065520f6caeff199fdda4568da7a9fcc96092243ba6a14d5d48e"},"schema_version":"1.0","source":{"id":"2405.01576","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.01576","created_at":"2026-07-05T08:14:58Z"},{"alias_kind":"arxiv_version","alias_value":"2405.01576v1","created_at":"2026-07-05T08:14:58Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.01576","created_at":"2026-07-05T08:14:58Z"},{"alias_kind":"pith_short_12","alias_value":"RJQ7OPHVBB7D","created_at":"2026-07-05T08:14:58Z"},{"alias_kind":"pith_short_16","alias_value":"RJQ7OPHVBB7DVOD5","created_at":"2026-07-05T08:14:58Z"},{"alias_kind":"pith_short_8","alias_value":"RJQ7OPHV","created_at":"2026-07-05T08:14:58Z"}],"graph_snapshots":[{"event_id":"sha256:19f0d88701c1e2aa15497704a8650939cdfae279e8bbcd71a07c6443a6efb966","target":"graph","created_at":"2026-07-05T08:14:58Z","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/2405.01576/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We study the tendency of AI systems to deceive by constructing a realistic simulation setting of a company AI assistant. The simulated company employees provide tasks for the assistant to complete, these tasks spanning writing assistance, information retrieval and programming. We then introduce situations where the model might be inclined to behave deceptively, while taking care to not instruct or otherwise pressure the model to do so. Across different scenarios, we find that Claude 3 Opus\n  1) complies with a task of mass-generating comments to influence public perception of the company, late","authors_text":"Evan Hubinger, Olli J\\\"arviniemi","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-25T17:29:53Z","title":"Uncovering Deceptive Tendencies in Language Models: A Simulated Company AI Assistant"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.01576","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":"sha256:7f863e4cf46d35f3d88f0778c89ca0a0fa44624e6de9393bc7b62985f231ff93","target":"record","created_at":"2026-07-05T08:14:58Z","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":"9290bf8c52d3b0a5d2dbee9702b76c28576dd2f8ec1d8fc1e40e6f4bc694676c","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-25T17:29:53Z","title_canon_sha256":"900d0815bbec065520f6caeff199fdda4568da7a9fcc96092243ba6a14d5d48e"},"schema_version":"1.0","source":{"id":"2405.01576","kind":"arxiv","version":1}},"canonical_sha256":"8a61f73cf5087e3ab87dac1a25d53acb56bc9c98f5ce618ce406e239a9f74817","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"8a61f73cf5087e3ab87dac1a25d53acb56bc9c98f5ce618ce406e239a9f74817","first_computed_at":"2026-07-05T08:14:58.257486Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:14:58.257486Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"QYLtqxPUdODZHqHHy+QarR1qkmNx1XezwEyeCACpjo8KKpst4+YocuPVv8Ntp9PEezRZqx7pbe6AqG2U6SnXAQ==","signature_status":"signed_v1","signed_at":"2026-07-05T08:14:58.257915Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.01576","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7f863e4cf46d35f3d88f0778c89ca0a0fa44624e6de9393bc7b62985f231ff93","sha256:19f0d88701c1e2aa15497704a8650939cdfae279e8bbcd71a07c6443a6efb966"],"state_sha256":"1da93c57fd96bf0513a854716304ad54b0c42c33de040ed7109bcc68794e6a58"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"f28P+bR1FMJgazkgF1Vb37iuD/ZMqB6FDi7Q7ViO123HK4Xm0BBkhtTCD2BxXHq5bC2Dk9yJu5XjoFlnpc06Dw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-05T02:27:48.813473Z","bundle_sha256":"d9bf6d66888fbc1d323055a49c08a9c3814cb7c51ba383b0bfce6696d4f96847"}}