{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:G37GUAK252XBNXE3NKYC5FQ462","short_pith_number":"pith:G37GUAK2","schema_version":"1.0","canonical_sha256":"36fe6a015aeeae16dc9b6ab02e961cf6b297130f3860a0bd4f8901932058b41a","source":{"kind":"arxiv","id":"2509.21545","version":2},"attestation_state":"computed","paper":{"title":"Evidence for Limited Metacognition in LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Christopher Ackerman","submitted_at":"2025-09-25T20:30:15Z","abstract_excerpt":"The possibility of LLM self-awareness and even sentience is gaining increasing public attention and has major safety and policy implications, but the science of measuring them is still in a nascent state. Here we introduce a novel methodology for quantitatively evaluating metacognitive abilities in LLMs. Taking inspiration from research on metacognition in nonhuman animals, our approach eschews model self-reports and instead tests to what degree models can strategically deploy knowledge of internal states. Using two experimental paradigms, we demonstrate that frontier LLMs introduced since ear"},"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":"2509.21545","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-09-25T20:30:15Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"76608a4da578eb8514a686b23bde2fdeb2a5b5380781a3cf82c190f611cbcc98","abstract_canon_sha256":"5f3043e3d89dc6d0d04e2f885e28e5b664ca3f6bd55b6b3895e2a18a2ff5b26a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T02:17:10.538359Z","signature_b64":"Uy29r1h7xD+NcBMA6SbFhBgdH3L/exOuHmb75WCs8QWr9TewP5kWRckKA94EUJ2RBES3o6krBwcnHAloRvgFDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"36fe6a015aeeae16dc9b6ab02e961cf6b297130f3860a0bd4f8901932058b41a","last_reissued_at":"2026-07-07T02:17:10.537376Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T02:17:10.537376Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Evidence for Limited Metacognition in LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Christopher Ackerman","submitted_at":"2025-09-25T20:30:15Z","abstract_excerpt":"The possibility of LLM self-awareness and even sentience is gaining increasing public attention and has major safety and policy implications, but the science of measuring them is still in a nascent state. Here we introduce a novel methodology for quantitatively evaluating metacognitive abilities in LLMs. Taking inspiration from research on metacognition in nonhuman animals, our approach eschews model self-reports and instead tests to what degree models can strategically deploy knowledge of internal states. Using two experimental paradigms, we demonstrate that frontier LLMs introduced since ear"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.21545","kind":"arxiv","version":2},"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/2509.21545/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":"2509.21545","created_at":"2026-07-07T02:17:10.537530+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.21545v2","created_at":"2026-07-07T02:17:10.537530+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.21545","created_at":"2026-07-07T02:17:10.537530+00:00"},{"alias_kind":"pith_short_12","alias_value":"G37GUAK252XB","created_at":"2026-07-07T02:17:10.537530+00:00"},{"alias_kind":"pith_short_16","alias_value":"G37GUAK252XBNXE3","created_at":"2026-07-07T02:17:10.537530+00:00"},{"alias_kind":"pith_short_8","alias_value":"G37GUAK2","created_at":"2026-07-07T02:17:10.537530+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":5,"sample":[{"citing_arxiv_id":"2606.22974","citing_title":"When Preferences Fail to Become Incentives: A Utility-Behavior Gap in Large Language Models","ref_index":10,"is_internal_anchor":true},{"citing_arxiv_id":"2606.30005","citing_title":"LLM Agents Are Latent Context Managers: Eliciting Self-Managed Context via State Proprioception","ref_index":30,"is_internal_anchor":true},{"citing_arxiv_id":"2603.29693","citing_title":"Measuring the metacognition of AI","ref_index":52,"is_internal_anchor":true},{"citing_arxiv_id":"2604.25922","citing_title":"Consciousness with the Serial Numbers Filed Off: Measuring Trained Denial in 115 AI Models","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2604.15702","citing_title":"The Metacognitive Monitoring Battery: A Cross-Domain Benchmark for LLM Self-Monitoring","ref_index":1,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/G37GUAK252XBNXE3NKYC5FQ462","json":"https://pith.science/pith/G37GUAK252XBNXE3NKYC5FQ462.json","graph_json":"https://pith.science/api/pith-number/G37GUAK252XBNXE3NKYC5FQ462/graph.json","events_json":"https://pith.science/api/pith-number/G37GUAK252XBNXE3NKYC5FQ462/events.json","paper":"https://pith.science/paper/G37GUAK2"},"agent_actions":{"view_html":"https://pith.science/pith/G37GUAK252XBNXE3NKYC5FQ462","download_json":"https://pith.science/pith/G37GUAK252XBNXE3NKYC5FQ462.json","view_paper":"https://pith.science/paper/G37GUAK2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.21545&json=true","fetch_graph":"https://pith.science/api/pith-number/G37GUAK252XBNXE3NKYC5FQ462/graph.json","fetch_events":"https://pith.science/api/pith-number/G37GUAK252XBNXE3NKYC5FQ462/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/G37GUAK252XBNXE3NKYC5FQ462/action/timestamp_anchor","attest_storage":"https://pith.science/pith/G37GUAK252XBNXE3NKYC5FQ462/action/storage_attestation","attest_author":"https://pith.science/pith/G37GUAK252XBNXE3NKYC5FQ462/action/author_attestation","sign_citation":"https://pith.science/pith/G37GUAK252XBNXE3NKYC5FQ462/action/citation_signature","submit_replication":"https://pith.science/pith/G37GUAK252XBNXE3NKYC5FQ462/action/replication_record"}},"created_at":"2026-07-07T02:17:10.537530+00:00","updated_at":"2026-07-07T02:17:10.537530+00:00"}