{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:SBCFSPUKDPRPLG4K4EHQUZIWTS","short_pith_number":"pith:SBCFSPUK","schema_version":"1.0","canonical_sha256":"9044593e8a1be2f59b8ae10f0a65169c97ace16723d77590159ecf5c01f27009","source":{"kind":"arxiv","id":"2402.18496","version":3},"attestation_state":"computed","paper":{"title":"Language Models Represent Beliefs of Self and Others","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Wentao Zhu, Yizhou Wang, Zhining Zhang","submitted_at":"2024-02-28T17:25:59Z","abstract_excerpt":"Understanding and attributing mental states, known as Theory of Mind (ToM), emerges as a fundamental capability for human social reasoning. While Large Language Models (LLMs) appear to possess certain ToM abilities, the mechanisms underlying these capabilities remain elusive. In this study, we discover that it is possible to linearly decode the belief status from the perspectives of various agents through neural activations of language models, indicating the existence of internal representations of self and others' beliefs. By manipulating these representations, we observe dramatic changes in "},"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":"2402.18496","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2024-02-28T17:25:59Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"f92ef303dc6aecc9a11a489d5fba9473e012ffd62957a0185c17e3f194602814","abstract_canon_sha256":"c47edf41cf89d10b2fd46ca2439b5fefef55737305585cb06ac1f5798808b9ea"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:25:01.284538Z","signature_b64":"LSTgGGVpeacvkh1iXvUftB9dm+hieqg3Ro4HJIMjlGPE0+WqyisLopt2tdraTzm3FvJ+cJbjUhM0Fu/G8beUCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9044593e8a1be2f59b8ae10f0a65169c97ace16723d77590159ecf5c01f27009","last_reissued_at":"2026-07-05T08:25:01.284123Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:25:01.284123Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Language Models Represent Beliefs of Self and Others","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Wentao Zhu, Yizhou Wang, Zhining Zhang","submitted_at":"2024-02-28T17:25:59Z","abstract_excerpt":"Understanding and attributing mental states, known as Theory of Mind (ToM), emerges as a fundamental capability for human social reasoning. While Large Language Models (LLMs) appear to possess certain ToM abilities, the mechanisms underlying these capabilities remain elusive. In this study, we discover that it is possible to linearly decode the belief status from the perspectives of various agents through neural activations of language models, indicating the existence of internal representations of self and others' beliefs. By manipulating these representations, we observe dramatic changes in "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.18496","kind":"arxiv","version":3},"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/2402.18496/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":"2402.18496","created_at":"2026-07-05T08:25:01.284175+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.18496v3","created_at":"2026-07-05T08:25:01.284175+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.18496","created_at":"2026-07-05T08:25:01.284175+00:00"},{"alias_kind":"pith_short_12","alias_value":"SBCFSPUKDPRP","created_at":"2026-07-05T08:25:01.284175+00:00"},{"alias_kind":"pith_short_16","alias_value":"SBCFSPUKDPRPLG4K","created_at":"2026-07-05T08:25:01.284175+00:00"},{"alias_kind":"pith_short_8","alias_value":"SBCFSPUK","created_at":"2026-07-05T08:25:01.284175+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.18284","citing_title":"Breaking the Solver Bottleneck: Training Task Generators at the Learnable Frontier","ref_index":113,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12412","citing_title":"Stories in Space: In-Context Learning Trajectories in Conceptual Belief Space","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03255","citing_title":"Do LLMs have core beliefs?","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07729","citing_title":"Emotion Concepts and their Function in a Large Language Model","ref_index":39,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SBCFSPUKDPRPLG4K4EHQUZIWTS","json":"https://pith.science/pith/SBCFSPUKDPRPLG4K4EHQUZIWTS.json","graph_json":"https://pith.science/api/pith-number/SBCFSPUKDPRPLG4K4EHQUZIWTS/graph.json","events_json":"https://pith.science/api/pith-number/SBCFSPUKDPRPLG4K4EHQUZIWTS/events.json","paper":"https://pith.science/paper/SBCFSPUK"},"agent_actions":{"view_html":"https://pith.science/pith/SBCFSPUKDPRPLG4K4EHQUZIWTS","download_json":"https://pith.science/pith/SBCFSPUKDPRPLG4K4EHQUZIWTS.json","view_paper":"https://pith.science/paper/SBCFSPUK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.18496&json=true","fetch_graph":"https://pith.science/api/pith-number/SBCFSPUKDPRPLG4K4EHQUZIWTS/graph.json","fetch_events":"https://pith.science/api/pith-number/SBCFSPUKDPRPLG4K4EHQUZIWTS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SBCFSPUKDPRPLG4K4EHQUZIWTS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SBCFSPUKDPRPLG4K4EHQUZIWTS/action/storage_attestation","attest_author":"https://pith.science/pith/SBCFSPUKDPRPLG4K4EHQUZIWTS/action/author_attestation","sign_citation":"https://pith.science/pith/SBCFSPUKDPRPLG4K4EHQUZIWTS/action/citation_signature","submit_replication":"https://pith.science/pith/SBCFSPUKDPRPLG4K4EHQUZIWTS/action/replication_record"}},"created_at":"2026-07-05T08:25:01.284175+00:00","updated_at":"2026-07-05T08:25:01.284175+00:00"}