{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:PV6ERVSMHBKP25BDIFCPW4FPL4","short_pith_number":"pith:PV6ERVSM","schema_version":"1.0","canonical_sha256":"7d7c48d64c3854fd74234144fb70af5f17b8d287594242bf64fb8b5ddac968d8","source":{"kind":"arxiv","id":"2412.07446","version":4},"attestation_state":"computed","paper":{"title":"A Causal World Model Underlying Next Token Prediction: Exploring GPT in a Controlled Environment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG","stat.ML"],"primary_cat":"cs.AI","authors_text":"Estelle Aflalo, Raanan Y. Rohekar, Sungduk Yu, Vasudev Lal, Yaniv Gurwicz","submitted_at":"2024-12-10T12:05:03Z","abstract_excerpt":"Are generative pre-trained transformer (GPT) models, trained only to predict the next token, implicitly learning a world model from which sequences are generated one token at a time? We address this question by deriving a causal interpretation of the attention mechanism in GPT and presenting a causal world model that arises from this interpretation. Furthermore, we propose that GPT models, at inference time, can be utilized for zero-shot causal structure learning for input sequences, and introduce a corresponding confidence score. Empirical tests were conducted in controlled environments using"},"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":"2412.07446","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-12-10T12:05:03Z","cross_cats_sorted":["cs.CL","cs.LG","stat.ML"],"title_canon_sha256":"d60b09e5ade8a0e2fc40adf67a1d4ea33cf4cfa7aeb72a7bbb2cd5da4a03d3b6","abstract_canon_sha256":"564098f430b9328a5b846abc456f5e5fac56aacf45f4d81d22b56e7faa6e0edc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:32:21.659935Z","signature_b64":"UAxUp6Q+Xqz+H9oVzycNmMjk9kEk7rbK9Pyjd8pBesUeZOvpOVt95cfsBJLmsPFvr3IDvGijQ7hvwmHaAj6/AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7d7c48d64c3854fd74234144fb70af5f17b8d287594242bf64fb8b5ddac968d8","last_reissued_at":"2026-07-05T11:32:21.659422Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:32:21.659422Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Causal World Model Underlying Next Token Prediction: Exploring GPT in a Controlled Environment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG","stat.ML"],"primary_cat":"cs.AI","authors_text":"Estelle Aflalo, Raanan Y. Rohekar, Sungduk Yu, Vasudev Lal, Yaniv Gurwicz","submitted_at":"2024-12-10T12:05:03Z","abstract_excerpt":"Are generative pre-trained transformer (GPT) models, trained only to predict the next token, implicitly learning a world model from which sequences are generated one token at a time? We address this question by deriving a causal interpretation of the attention mechanism in GPT and presenting a causal world model that arises from this interpretation. Furthermore, we propose that GPT models, at inference time, can be utilized for zero-shot causal structure learning for input sequences, and introduce a corresponding confidence score. Empirical tests were conducted in controlled environments using"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.07446","kind":"arxiv","version":4},"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/2412.07446/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":"2412.07446","created_at":"2026-07-05T11:32:21.659488+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.07446v4","created_at":"2026-07-05T11:32:21.659488+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.07446","created_at":"2026-07-05T11:32:21.659488+00:00"},{"alias_kind":"pith_short_12","alias_value":"PV6ERVSMHBKP","created_at":"2026-07-05T11:32:21.659488+00:00"},{"alias_kind":"pith_short_16","alias_value":"PV6ERVSMHBKP25BD","created_at":"2026-07-05T11:32:21.659488+00:00"},{"alias_kind":"pith_short_8","alias_value":"PV6ERVSM","created_at":"2026-07-05T11:32:21.659488+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PV6ERVSMHBKP25BDIFCPW4FPL4","json":"https://pith.science/pith/PV6ERVSMHBKP25BDIFCPW4FPL4.json","graph_json":"https://pith.science/api/pith-number/PV6ERVSMHBKP25BDIFCPW4FPL4/graph.json","events_json":"https://pith.science/api/pith-number/PV6ERVSMHBKP25BDIFCPW4FPL4/events.json","paper":"https://pith.science/paper/PV6ERVSM"},"agent_actions":{"view_html":"https://pith.science/pith/PV6ERVSMHBKP25BDIFCPW4FPL4","download_json":"https://pith.science/pith/PV6ERVSMHBKP25BDIFCPW4FPL4.json","view_paper":"https://pith.science/paper/PV6ERVSM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.07446&json=true","fetch_graph":"https://pith.science/api/pith-number/PV6ERVSMHBKP25BDIFCPW4FPL4/graph.json","fetch_events":"https://pith.science/api/pith-number/PV6ERVSMHBKP25BDIFCPW4FPL4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PV6ERVSMHBKP25BDIFCPW4FPL4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PV6ERVSMHBKP25BDIFCPW4FPL4/action/storage_attestation","attest_author":"https://pith.science/pith/PV6ERVSMHBKP25BDIFCPW4FPL4/action/author_attestation","sign_citation":"https://pith.science/pith/PV6ERVSMHBKP25BDIFCPW4FPL4/action/citation_signature","submit_replication":"https://pith.science/pith/PV6ERVSMHBKP25BDIFCPW4FPL4/action/replication_record"}},"created_at":"2026-07-05T11:32:21.659488+00:00","updated_at":"2026-07-05T11:32:21.659488+00:00"}