{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:VA4Y7V4PFUACVAOAPUSWDD7B25","short_pith_number":"pith:VA4Y7V4P","schema_version":"1.0","canonical_sha256":"a8398fd78f2d002a81c07d25618fe1d76818797372b66fda54aa9ca73cb8427c","source":{"kind":"arxiv","id":"2301.13442","version":2},"attestation_state":"computed","paper":{"title":"Scaling laws for single-agent reinforcement learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Jacob Hilton, Jie Tang, John Schulman","submitted_at":"2023-01-31T06:38:53Z","abstract_excerpt":"Recent work has shown that, in generative modeling, cross-entropy loss improves smoothly with model size and training compute, following a power law plus constant scaling law. One challenge in extending these results to reinforcement learning is that the main performance objective of interest, mean episode return, need not vary smoothly. To overcome this, we introduce *intrinsic performance*, a monotonic function of the return defined as the minimum compute required to achieve the given return across a family of models of different sizes. We find that, across a range of environments, intrinsic"},"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":"2301.13442","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-01-31T06:38:53Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"401b1d678a9a05a577be57e0fd7b1f7d56658be518e1d9155a6918acff41378e","abstract_canon_sha256":"735bb2cc758db04db3d0682322595dc27eb3476a343dd302e97cc7154a65a747"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:43:19.429686Z","signature_b64":"t02eHzcx4YwQnF0pDtIubLN3aftDhu/cFf2MYnodnVK6ZhcWbJAKyyI6Lb5PYitG1xWARgRPhYrdAt3R14C5Cw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a8398fd78f2d002a81c07d25618fe1d76818797372b66fda54aa9ca73cb8427c","last_reissued_at":"2026-07-05T05:43:19.429207Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:43:19.429207Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Scaling laws for single-agent reinforcement learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Jacob Hilton, Jie Tang, John Schulman","submitted_at":"2023-01-31T06:38:53Z","abstract_excerpt":"Recent work has shown that, in generative modeling, cross-entropy loss improves smoothly with model size and training compute, following a power law plus constant scaling law. One challenge in extending these results to reinforcement learning is that the main performance objective of interest, mean episode return, need not vary smoothly. To overcome this, we introduce *intrinsic performance*, a monotonic function of the return defined as the minimum compute required to achieve the given return across a family of models of different sizes. We find that, across a range of environments, intrinsic"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2301.13442","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/2301.13442/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":"2301.13442","created_at":"2026-07-05T05:43:19.429264+00:00"},{"alias_kind":"arxiv_version","alias_value":"2301.13442v2","created_at":"2026-07-05T05:43:19.429264+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2301.13442","created_at":"2026-07-05T05:43:19.429264+00:00"},{"alias_kind":"pith_short_12","alias_value":"VA4Y7V4PFUAC","created_at":"2026-07-05T05:43:19.429264+00:00"},{"alias_kind":"pith_short_16","alias_value":"VA4Y7V4PFUACVAOA","created_at":"2026-07-05T05:43:19.429264+00:00"},{"alias_kind":"pith_short_8","alias_value":"VA4Y7V4P","created_at":"2026-07-05T05:43:19.429264+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":8,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2607.01490","citing_title":"Don't Let Gains FADE: Breaking Down Policy Gradient Weights in RL","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2607.00913","citing_title":"Two AI Metrics Diverged: Will it Make All the Difference?","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02794","citing_title":"Scaling Laws for Neural-Network Quantum States","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.26248","citing_title":"Unified Neural Scaling Laws","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2509.24244","citing_title":"Model Merging Scaling Laws in Large Language Models","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2309.16797","citing_title":"Promptbreeder: Self-Referential Self-Improvement Via Prompt Evolution","ref_index":148,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06732","citing_title":"On Training in Imagination","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06732","citing_title":"On Training in Imagination","ref_index":27,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VA4Y7V4PFUACVAOAPUSWDD7B25","json":"https://pith.science/pith/VA4Y7V4PFUACVAOAPUSWDD7B25.json","graph_json":"https://pith.science/api/pith-number/VA4Y7V4PFUACVAOAPUSWDD7B25/graph.json","events_json":"https://pith.science/api/pith-number/VA4Y7V4PFUACVAOAPUSWDD7B25/events.json","paper":"https://pith.science/paper/VA4Y7V4P"},"agent_actions":{"view_html":"https://pith.science/pith/VA4Y7V4PFUACVAOAPUSWDD7B25","download_json":"https://pith.science/pith/VA4Y7V4PFUACVAOAPUSWDD7B25.json","view_paper":"https://pith.science/paper/VA4Y7V4P","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2301.13442&json=true","fetch_graph":"https://pith.science/api/pith-number/VA4Y7V4PFUACVAOAPUSWDD7B25/graph.json","fetch_events":"https://pith.science/api/pith-number/VA4Y7V4PFUACVAOAPUSWDD7B25/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VA4Y7V4PFUACVAOAPUSWDD7B25/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VA4Y7V4PFUACVAOAPUSWDD7B25/action/storage_attestation","attest_author":"https://pith.science/pith/VA4Y7V4PFUACVAOAPUSWDD7B25/action/author_attestation","sign_citation":"https://pith.science/pith/VA4Y7V4PFUACVAOAPUSWDD7B25/action/citation_signature","submit_replication":"https://pith.science/pith/VA4Y7V4PFUACVAOAPUSWDD7B25/action/replication_record"}},"created_at":"2026-07-05T05:43:19.429264+00:00","updated_at":"2026-07-05T05:43:19.429264+00:00"}