{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:JWAFCZSEBMIFUW4Y7ZS3ZH6TST","short_pith_number":"pith:JWAFCZSE","schema_version":"1.0","canonical_sha256":"4d805166440b105a5b98fe65bc9fd394e67abe32b6284c7eeb1a9f26a04d6feb","source":{"kind":"arxiv","id":"2306.14892","version":1},"attestation_state":"computed","paper":{"title":"Supervised Pretraining Can Learn In-Context Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Aldo Pacchiano, Annie Xie, Chelsea Finn, Emma Brunskill, Jonathan N. Lee, Ofir Nachum, Yash Chandak","submitted_at":"2023-06-26T17:58:50Z","abstract_excerpt":"Large transformer models trained on diverse datasets have shown a remarkable ability to learn in-context, achieving high few-shot performance on tasks they were not explicitly trained to solve. In this paper, we study the in-context learning capabilities of transformers in decision-making problems, i.e., reinforcement learning (RL) for bandits and Markov decision processes. To do so, we introduce and study Decision-Pretrained Transformer (DPT), a supervised pretraining method where the transformer predicts an optimal action given a query state and an in-context dataset of interactions, across "},"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":"2306.14892","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-06-26T17:58:50Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"df8f95ad9314113c76fe26118132ac6668422e6dec9863d8f3e774e11251d310","abstract_canon_sha256":"94139ee34686e8d902b4a5cb462ed76d14efbd4b269ed8c4f446b0d510e29a43"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:24:41.809670Z","signature_b64":"mc+cNPDxVn8xayBfv2Vw3XVuLzWkp8FYTZHYIIX1mSUOGZSAvSy5EFwM+Vv60QHR5vckcVLYt06wT/+UgSEgBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4d805166440b105a5b98fe65bc9fd394e67abe32b6284c7eeb1a9f26a04d6feb","last_reissued_at":"2026-07-05T06:24:41.809294Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:24:41.809294Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Supervised Pretraining Can Learn In-Context Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Aldo Pacchiano, Annie Xie, Chelsea Finn, Emma Brunskill, Jonathan N. Lee, Ofir Nachum, Yash Chandak","submitted_at":"2023-06-26T17:58:50Z","abstract_excerpt":"Large transformer models trained on diverse datasets have shown a remarkable ability to learn in-context, achieving high few-shot performance on tasks they were not explicitly trained to solve. In this paper, we study the in-context learning capabilities of transformers in decision-making problems, i.e., reinforcement learning (RL) for bandits and Markov decision processes. To do so, we introduce and study Decision-Pretrained Transformer (DPT), a supervised pretraining method where the transformer predicts an optimal action given a query state and an in-context dataset of interactions, across "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2306.14892","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/2306.14892/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":"2306.14892","created_at":"2026-07-05T06:24:41.809356+00:00"},{"alias_kind":"arxiv_version","alias_value":"2306.14892v1","created_at":"2026-07-05T06:24:41.809356+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2306.14892","created_at":"2026-07-05T06:24:41.809356+00:00"},{"alias_kind":"pith_short_12","alias_value":"JWAFCZSEBMIF","created_at":"2026-07-05T06:24:41.809356+00:00"},{"alias_kind":"pith_short_16","alias_value":"JWAFCZSEBMIFUW4Y","created_at":"2026-07-05T06:24:41.809356+00:00"},{"alias_kind":"pith_short_8","alias_value":"JWAFCZSE","created_at":"2026-07-05T06:24:41.809356+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.24962","citing_title":"Towards Scalable Multi-Task Reinforcement Learning with Large Decision Models","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2606.18812","citing_title":"Reinforcement Learning Foundation Models Should Already Be A Thing","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2309.16797","citing_title":"Promptbreeder: Self-Referential Self-Improvement Via Prompt Evolution","ref_index":194,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12288","citing_title":"TokenRatio: Principled Token-Level Preference Optimization via Ratio Matching","ref_index":82,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12288","citing_title":"TokenRatio: Principled Token-Level Preference Optimization via Ratio Matching","ref_index":82,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09727","citing_title":"One for All: A Non-Linear Transformer can Enable Cross-Domain Generalization for In-Context Reinforcement Learning","ref_index":9,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JWAFCZSEBMIFUW4Y7ZS3ZH6TST","json":"https://pith.science/pith/JWAFCZSEBMIFUW4Y7ZS3ZH6TST.json","graph_json":"https://pith.science/api/pith-number/JWAFCZSEBMIFUW4Y7ZS3ZH6TST/graph.json","events_json":"https://pith.science/api/pith-number/JWAFCZSEBMIFUW4Y7ZS3ZH6TST/events.json","paper":"https://pith.science/paper/JWAFCZSE"},"agent_actions":{"view_html":"https://pith.science/pith/JWAFCZSEBMIFUW4Y7ZS3ZH6TST","download_json":"https://pith.science/pith/JWAFCZSEBMIFUW4Y7ZS3ZH6TST.json","view_paper":"https://pith.science/paper/JWAFCZSE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2306.14892&json=true","fetch_graph":"https://pith.science/api/pith-number/JWAFCZSEBMIFUW4Y7ZS3ZH6TST/graph.json","fetch_events":"https://pith.science/api/pith-number/JWAFCZSEBMIFUW4Y7ZS3ZH6TST/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JWAFCZSEBMIFUW4Y7ZS3ZH6TST/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JWAFCZSEBMIFUW4Y7ZS3ZH6TST/action/storage_attestation","attest_author":"https://pith.science/pith/JWAFCZSEBMIFUW4Y7ZS3ZH6TST/action/author_attestation","sign_citation":"https://pith.science/pith/JWAFCZSEBMIFUW4Y7ZS3ZH6TST/action/citation_signature","submit_replication":"https://pith.science/pith/JWAFCZSEBMIFUW4Y7ZS3ZH6TST/action/replication_record"}},"created_at":"2026-07-05T06:24:41.809356+00:00","updated_at":"2026-07-05T06:24:41.809356+00:00"}