{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:2FHP6L5NQRFNAY6QXWPXPWLKWW","short_pith_number":"pith:2FHP6L5N","canonical_record":{"source":{"id":"2405.12999","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-10T23:24:18Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"21cb1a1727d6d57d30e454bff2e35929eeaa1da2e193560c83528105e7a4b8bc","abstract_canon_sha256":"7ddd325e10c13b772589bf25bb3caa4dcd619f8ad17c5f32ad685d84c3d7954f"},"schema_version":"1.0"},"canonical_sha256":"d14eff2fad844ad063d0bd9f77d96ab5bb90b694d6ae6f497764ee9718b9089c","source":{"kind":"arxiv","id":"2405.12999","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.12999","created_at":"2026-07-05T08:21:29Z"},{"alias_kind":"arxiv_version","alias_value":"2405.12999v1","created_at":"2026-07-05T08:21:29Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.12999","created_at":"2026-07-05T08:21:29Z"},{"alias_kind":"pith_short_12","alias_value":"2FHP6L5NQRFN","created_at":"2026-07-05T08:21:29Z"},{"alias_kind":"pith_short_16","alias_value":"2FHP6L5NQRFNAY6Q","created_at":"2026-07-05T08:21:29Z"},{"alias_kind":"pith_short_8","alias_value":"2FHP6L5N","created_at":"2026-07-05T08:21:29Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:2FHP6L5NQRFNAY6QXWPXPWLKWW","target":"record","payload":{"canonical_record":{"source":{"id":"2405.12999","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-10T23:24:18Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"21cb1a1727d6d57d30e454bff2e35929eeaa1da2e193560c83528105e7a4b8bc","abstract_canon_sha256":"7ddd325e10c13b772589bf25bb3caa4dcd619f8ad17c5f32ad685d84c3d7954f"},"schema_version":"1.0"},"canonical_sha256":"d14eff2fad844ad063d0bd9f77d96ab5bb90b694d6ae6f497764ee9718b9089c","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:21:29.807545Z","signature_b64":"C6WWWZanzP+0wTrIt91Lw3oWDGz505UOKMf2AHjIfdSSrOP3/urPsun+aEe2XVQihZTMizv4c5GaaQ9kk1vsDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d14eff2fad844ad063d0bd9f77d96ab5bb90b694d6ae6f497764ee9718b9089c","last_reissued_at":"2026-07-05T08:21:29.807076Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:21:29.807076Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.12999","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":"2026-07-05T08:21:29Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"t0Lep4aKYVPFgqrqQO3aN7OgnrAoAdt6Rpu1TaCykdgroa7R/Gjx2m8vb+pM9bo4q7oeR+gpRQFXY49ts9rrAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T20:40:21.000800Z"},"content_sha256":"3465651df6803e1f5529be4266c904abadef1b6bb51592c9ea3d59a6c33fad8b","schema_version":"1.0","event_id":"sha256:3465651df6803e1f5529be4266c904abadef1b6bb51592c9ea3d59a6c33fad8b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:2FHP6L5NQRFNAY6QXWPXPWLKWW","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"An Assessment of Model-On-Model Deception","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Julius Heitkoetter, Laker Newhouse, Michael Gerovitch","submitted_at":"2024-05-10T23:24:18Z","abstract_excerpt":"The trustworthiness of highly capable language models is put at risk when they are able to produce deceptive outputs. Moreover, when models are vulnerable to deception it undermines reliability. In this paper, we introduce a method to investigate complex, model-on-model deceptive scenarios. We create a dataset of over 10,000 misleading explanations by asking Llama-2 7B, 13B, 70B, and GPT-3.5 to justify the wrong answer for questions in the MMLU. We find that, when models read these explanations, they are all significantly deceived. Worryingly, models of all capabilities are successful at misle"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.12999","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/2405.12999/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":"2026-07-05T08:21:29Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"NMRFF3jaAlHk0VpXKFpsQa0zk/6Yfu+WrTVOw2tflJxTnNrPSZPlpUpNPyyz5jUz5RUh++1BN60pnFqMXRPqDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T20:40:21.001420Z"},"content_sha256":"952d3d745fc6f1cfcf7f7d8618dd0cf050b44b522db93bd5eca131850eb25293","schema_version":"1.0","event_id":"sha256:952d3d745fc6f1cfcf7f7d8618dd0cf050b44b522db93bd5eca131850eb25293"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/2FHP6L5NQRFNAY6QXWPXPWLKWW/bundle.json","state_url":"https://pith.science/pith/2FHP6L5NQRFNAY6QXWPXPWLKWW/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/2FHP6L5NQRFNAY6QXWPXPWLKWW/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":"2026-08-09T20:40:21Z","links":{"resolver":"https://pith.science/pith/2FHP6L5NQRFNAY6QXWPXPWLKWW","bundle":"https://pith.science/pith/2FHP6L5NQRFNAY6QXWPXPWLKWW/bundle.json","state":"https://pith.science/pith/2FHP6L5NQRFNAY6QXWPXPWLKWW/state.json","well_known_bundle":"https://pith.science/.well-known/pith/2FHP6L5NQRFNAY6QXWPXPWLKWW/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:2FHP6L5NQRFNAY6QXWPXPWLKWW","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":"7ddd325e10c13b772589bf25bb3caa4dcd619f8ad17c5f32ad685d84c3d7954f","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-10T23:24:18Z","title_canon_sha256":"21cb1a1727d6d57d30e454bff2e35929eeaa1da2e193560c83528105e7a4b8bc"},"schema_version":"1.0","source":{"id":"2405.12999","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.12999","created_at":"2026-07-05T08:21:29Z"},{"alias_kind":"arxiv_version","alias_value":"2405.12999v1","created_at":"2026-07-05T08:21:29Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.12999","created_at":"2026-07-05T08:21:29Z"},{"alias_kind":"pith_short_12","alias_value":"2FHP6L5NQRFN","created_at":"2026-07-05T08:21:29Z"},{"alias_kind":"pith_short_16","alias_value":"2FHP6L5NQRFNAY6Q","created_at":"2026-07-05T08:21:29Z"},{"alias_kind":"pith_short_8","alias_value":"2FHP6L5N","created_at":"2026-07-05T08:21:29Z"}],"graph_snapshots":[{"event_id":"sha256:952d3d745fc6f1cfcf7f7d8618dd0cf050b44b522db93bd5eca131850eb25293","target":"graph","created_at":"2026-07-05T08:21:29Z","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/2405.12999/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The trustworthiness of highly capable language models is put at risk when they are able to produce deceptive outputs. Moreover, when models are vulnerable to deception it undermines reliability. In this paper, we introduce a method to investigate complex, model-on-model deceptive scenarios. We create a dataset of over 10,000 misleading explanations by asking Llama-2 7B, 13B, 70B, and GPT-3.5 to justify the wrong answer for questions in the MMLU. We find that, when models read these explanations, they are all significantly deceived. Worryingly, models of all capabilities are successful at misle","authors_text":"Julius Heitkoetter, Laker Newhouse, Michael Gerovitch","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-10T23:24:18Z","title":"An Assessment of Model-On-Model Deception"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.12999","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":"sha256:3465651df6803e1f5529be4266c904abadef1b6bb51592c9ea3d59a6c33fad8b","target":"record","created_at":"2026-07-05T08:21:29Z","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":"7ddd325e10c13b772589bf25bb3caa4dcd619f8ad17c5f32ad685d84c3d7954f","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-10T23:24:18Z","title_canon_sha256":"21cb1a1727d6d57d30e454bff2e35929eeaa1da2e193560c83528105e7a4b8bc"},"schema_version":"1.0","source":{"id":"2405.12999","kind":"arxiv","version":1}},"canonical_sha256":"d14eff2fad844ad063d0bd9f77d96ab5bb90b694d6ae6f497764ee9718b9089c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"d14eff2fad844ad063d0bd9f77d96ab5bb90b694d6ae6f497764ee9718b9089c","first_computed_at":"2026-07-05T08:21:29.807076Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:21:29.807076Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"C6WWWZanzP+0wTrIt91Lw3oWDGz505UOKMf2AHjIfdSSrOP3/urPsun+aEe2XVQihZTMizv4c5GaaQ9kk1vsDw==","signature_status":"signed_v1","signed_at":"2026-07-05T08:21:29.807545Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.12999","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3465651df6803e1f5529be4266c904abadef1b6bb51592c9ea3d59a6c33fad8b","sha256:952d3d745fc6f1cfcf7f7d8618dd0cf050b44b522db93bd5eca131850eb25293"],"state_sha256":"aef569f9e9a9f2f1d9096512f4f97096f8a51504e4d909599515234eee6376ba"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"mD4JbWfoDl8PhVG+X2q/p6+YSJDAafIfVV+QLO6aXetFJ4A3ap3Gk9i2JkHQUJ6ydHOsND4IbVYhulQ02I1NDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T20:40:21.006559Z","bundle_sha256":"0d7ffbadd611c37404b8712b97fb73d21e23e7a1dc464762bc70c584c1c59928"}}