{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ZG3W2WIVLKO7MXVRV5BPJ6LHMR","short_pith_number":"pith:ZG3W2WIV","schema_version":"1.0","canonical_sha256":"c9b76d59155a9df65eb1af42f4f9676455d11b167c47f39ed9f5c307f106fcfe","source":{"kind":"arxiv","id":"2504.00285","version":1},"attestation_state":"computed","paper":{"title":"Do Large Language Models Exhibit Spontaneous Rational Deception?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Benjamin K. Bergen, Samuel M. Taylor","submitted_at":"2025-03-31T23:10:56Z","abstract_excerpt":"Large Language Models (LLMs) are effective at deceiving, when prompted to do so. But under what conditions do they deceive spontaneously? Models that demonstrate better performance on reasoning tasks are also better at prompted deception. Do they also increasingly deceive spontaneously in situations where it could be considered rational to do so? This study evaluates spontaneous deception produced by LLMs in a preregistered experimental protocol using tools from signaling theory. A range of proprietary closed-source and open-source LLMs are evaluated using modified 2x2 games (in the style of P"},"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":"2504.00285","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-03-31T23:10:56Z","cross_cats_sorted":[],"title_canon_sha256":"01c0c8961e7366fb141eb38c444ae1beb019f3f309f6c2266aa198da9d134bab","abstract_canon_sha256":"c9e148892791242754b2d8f581d24c4eb6e1fb535c75e2deec90aff8dc092b51"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:42:23.660221Z","signature_b64":"aNsZh0JTVp6u6bZol7Mi3Qct5bkI17GmRR0zT3huVweMDTwWxPjN3K0U0DR20LBfn6ZMb78KbBw4fn+lsfXrDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c9b76d59155a9df65eb1af42f4f9676455d11b167c47f39ed9f5c307f106fcfe","last_reissued_at":"2026-07-05T10:42:23.659767Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:42:23.659767Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Do Large Language Models Exhibit Spontaneous Rational Deception?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Benjamin K. Bergen, Samuel M. Taylor","submitted_at":"2025-03-31T23:10:56Z","abstract_excerpt":"Large Language Models (LLMs) are effective at deceiving, when prompted to do so. But under what conditions do they deceive spontaneously? Models that demonstrate better performance on reasoning tasks are also better at prompted deception. Do they also increasingly deceive spontaneously in situations where it could be considered rational to do so? This study evaluates spontaneous deception produced by LLMs in a preregistered experimental protocol using tools from signaling theory. A range of proprietary closed-source and open-source LLMs are evaluated using modified 2x2 games (in the style of P"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.00285","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/2504.00285/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":"2504.00285","created_at":"2026-07-05T10:42:23.659823+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.00285v1","created_at":"2026-07-05T10:42:23.659823+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.00285","created_at":"2026-07-05T10:42:23.659823+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZG3W2WIVLKO7","created_at":"2026-07-05T10:42:23.659823+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZG3W2WIVLKO7MXVR","created_at":"2026-07-05T10:42:23.659823+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZG3W2WIV","created_at":"2026-07-05T10:42:23.659823+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.13310","citing_title":"RogueAI: A Reverse Turing Test for Detecting Licensed AI Deception in Dialogue","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31916","citing_title":"Theory of Mind and Persuasion Beyond Conversation: Assessing the Capacity of LLMs to Induce Belief States via Planning and Action","ref_index":66,"is_internal_anchor":false},{"citing_arxiv_id":"2606.28456","citing_title":"Is Lying an Emergent Behaviour in LLMs? Evidence from Gaslighting AI agents in a Sustainability Game","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.19270","citing_title":"DECOR: Auditing LLM Deception via Information Manipulation Theory","ref_index":25,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZG3W2WIVLKO7MXVRV5BPJ6LHMR","json":"https://pith.science/pith/ZG3W2WIVLKO7MXVRV5BPJ6LHMR.json","graph_json":"https://pith.science/api/pith-number/ZG3W2WIVLKO7MXVRV5BPJ6LHMR/graph.json","events_json":"https://pith.science/api/pith-number/ZG3W2WIVLKO7MXVRV5BPJ6LHMR/events.json","paper":"https://pith.science/paper/ZG3W2WIV"},"agent_actions":{"view_html":"https://pith.science/pith/ZG3W2WIVLKO7MXVRV5BPJ6LHMR","download_json":"https://pith.science/pith/ZG3W2WIVLKO7MXVRV5BPJ6LHMR.json","view_paper":"https://pith.science/paper/ZG3W2WIV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.00285&json=true","fetch_graph":"https://pith.science/api/pith-number/ZG3W2WIVLKO7MXVRV5BPJ6LHMR/graph.json","fetch_events":"https://pith.science/api/pith-number/ZG3W2WIVLKO7MXVRV5BPJ6LHMR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZG3W2WIVLKO7MXVRV5BPJ6LHMR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZG3W2WIVLKO7MXVRV5BPJ6LHMR/action/storage_attestation","attest_author":"https://pith.science/pith/ZG3W2WIVLKO7MXVRV5BPJ6LHMR/action/author_attestation","sign_citation":"https://pith.science/pith/ZG3W2WIVLKO7MXVRV5BPJ6LHMR/action/citation_signature","submit_replication":"https://pith.science/pith/ZG3W2WIVLKO7MXVRV5BPJ6LHMR/action/replication_record"}},"created_at":"2026-07-05T10:42:23.659823+00:00","updated_at":"2026-07-05T10:42:23.659823+00:00"}