{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:FW6RX3KWYLHWIO3QFZ6HFK7JJ4","short_pith_number":"pith:FW6RX3KW","schema_version":"1.0","canonical_sha256":"2dbd1bed56c2cf643b702e7c72abe94f204ccb10ec6d46de33c217349723d032","source":{"kind":"arxiv","id":"2102.06701","version":2},"attestation_state":"computed","paper":{"title":"Explaining Neural Scaling Laws","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cond-mat.dis-nn","stat.ML"],"primary_cat":"cs.LG","authors_text":"Ethan Dyer, Jaehoon Lee, Jared Kaplan, Utkarsh Sharma, Yasaman Bahri","submitted_at":"2021-02-12T18:57:46Z","abstract_excerpt":"The population loss of trained deep neural networks often follows precise power-law scaling relations with either the size of the training dataset or the number of parameters in the network. We propose a theory that explains the origins of and connects these scaling laws. We identify variance-limited and resolution-limited scaling behavior for both dataset and model size, for a total of four scaling regimes. The variance-limited scaling follows simply from the existence of a well-behaved infinite data or infinite width limit, while the resolution-limited regime can be explained by positing tha"},"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":"2102.06701","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-02-12T18:57:46Z","cross_cats_sorted":["cond-mat.dis-nn","stat.ML"],"title_canon_sha256":"b1344dcd6025f41668dbbe41827c24452847de6e7a4d575bcd7642d478e11a47","abstract_canon_sha256":"26e02a56b65544bb5dd4635d4ae5accccbacf5bb57f90e18361cb0608b8a9c9a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:37:11.417630Z","signature_b64":"1eTnoSx/6ODqtW8Exp+nRyCUmOSFJn4u0XKh5litDXicEEbx1Fxq6N+tgzyDqKZ/uTonseyHmbzHmTLnDsBjAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2dbd1bed56c2cf643b702e7c72abe94f204ccb10ec6d46de33c217349723d032","last_reissued_at":"2026-07-05T08:37:11.417137Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:37:11.417137Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Explaining Neural Scaling Laws","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cond-mat.dis-nn","stat.ML"],"primary_cat":"cs.LG","authors_text":"Ethan Dyer, Jaehoon Lee, Jared Kaplan, Utkarsh Sharma, Yasaman Bahri","submitted_at":"2021-02-12T18:57:46Z","abstract_excerpt":"The population loss of trained deep neural networks often follows precise power-law scaling relations with either the size of the training dataset or the number of parameters in the network. We propose a theory that explains the origins of and connects these scaling laws. We identify variance-limited and resolution-limited scaling behavior for both dataset and model size, for a total of four scaling regimes. The variance-limited scaling follows simply from the existence of a well-behaved infinite data or infinite width limit, while the resolution-limited regime can be explained by positing tha"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2102.06701","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/2102.06701/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":"2102.06701","created_at":"2026-07-05T08:37:11.417197+00:00"},{"alias_kind":"arxiv_version","alias_value":"2102.06701v2","created_at":"2026-07-05T08:37:11.417197+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2102.06701","created_at":"2026-07-05T08:37:11.417197+00:00"},{"alias_kind":"pith_short_12","alias_value":"FW6RX3KWYLHW","created_at":"2026-07-05T08:37:11.417197+00:00"},{"alias_kind":"pith_short_16","alias_value":"FW6RX3KWYLHWIO3Q","created_at":"2026-07-05T08:37:11.417197+00:00"},{"alias_kind":"pith_short_8","alias_value":"FW6RX3KW","created_at":"2026-07-05T08:37:11.417197+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":17,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.20299","citing_title":"Statistical Properties of Training & Generalization","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02385","citing_title":"How Optimality Structures Sparse Dictionaries: A Theory for Understanding SAE Representations","ref_index":117,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31282","citing_title":"Revisiting the Volume Hypothesis","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00047","citing_title":"Comprehensive AI governance requires addressing non-model gains","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2606.21585","citing_title":"A Transport-Based Geometry of Belief-Cost","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.26248","citing_title":"Unified Neural Scaling Laws","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2606.20299","citing_title":"Statistical Properties of Training & Generalization","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.23591","citing_title":"Asymmetric Scaling Laws from Sparse Features","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07870","citing_title":"Spectral Dynamics in Deep Networks: Feature Learning, Outlier Escape, and Learning Rate Transfer","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20196","citing_title":"Data Scaling as Progressive Coverage of a Predictive Contribution Spectrum","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2210.10760","citing_title":"Scaling Laws for Reward Model Overoptimization","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2305.16264","citing_title":"Scaling Data-Constrained Language Models","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2405.07987","citing_title":"The Platonic Representation Hypothesis","ref_index":40,"is_internal_anchor":false},{"citing_arxiv_id":"2603.28964","citing_title":"Spectral Edge Dynamics: An Analytical-Empirical Study of Phase Transitions in Neural Network Training","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2404.19756","citing_title":"KAN: Kolmogorov-Arnold Networks","ref_index":77,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07870","citing_title":"Spectral Dynamics in Deep Networks: Feature Learning, Outlier Escape, and Learning Rate Transfer","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2604.16842","citing_title":"Singularity Formation: Synergy in Theoretical, Numerical and Machine Learning Approaches","ref_index":18,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FW6RX3KWYLHWIO3QFZ6HFK7JJ4","json":"https://pith.science/pith/FW6RX3KWYLHWIO3QFZ6HFK7JJ4.json","graph_json":"https://pith.science/api/pith-number/FW6RX3KWYLHWIO3QFZ6HFK7JJ4/graph.json","events_json":"https://pith.science/api/pith-number/FW6RX3KWYLHWIO3QFZ6HFK7JJ4/events.json","paper":"https://pith.science/paper/FW6RX3KW"},"agent_actions":{"view_html":"https://pith.science/pith/FW6RX3KWYLHWIO3QFZ6HFK7JJ4","download_json":"https://pith.science/pith/FW6RX3KWYLHWIO3QFZ6HFK7JJ4.json","view_paper":"https://pith.science/paper/FW6RX3KW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2102.06701&json=true","fetch_graph":"https://pith.science/api/pith-number/FW6RX3KWYLHWIO3QFZ6HFK7JJ4/graph.json","fetch_events":"https://pith.science/api/pith-number/FW6RX3KWYLHWIO3QFZ6HFK7JJ4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FW6RX3KWYLHWIO3QFZ6HFK7JJ4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FW6RX3KWYLHWIO3QFZ6HFK7JJ4/action/storage_attestation","attest_author":"https://pith.science/pith/FW6RX3KWYLHWIO3QFZ6HFK7JJ4/action/author_attestation","sign_citation":"https://pith.science/pith/FW6RX3KWYLHWIO3QFZ6HFK7JJ4/action/citation_signature","submit_replication":"https://pith.science/pith/FW6RX3KWYLHWIO3QFZ6HFK7JJ4/action/replication_record"}},"created_at":"2026-07-05T08:37:11.417197+00:00","updated_at":"2026-07-05T08:37:11.417197+00:00"}