{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:2KJ3OKTNIPY7M2MF7ALQMZRPJZ","short_pith_number":"pith:2KJ3OKTN","schema_version":"1.0","canonical_sha256":"d293b72a6d43f1f66985f81706662f4e580fc1791a896c8ac81b8cb834489220","source":{"kind":"arxiv","id":"2404.10102","version":2},"attestation_state":"computed","paper":{"title":"Chinchilla Scaling: A replication attempt","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Ege Erdil, Josh You, Matthew Barnett, Tamay Besiroglu","submitted_at":"2024-04-15T19:19:56Z","abstract_excerpt":"Hoffmann et al. (2022) propose three methods for estimating a compute-optimal scaling law. We attempt to replicate their third estimation procedure, which involves fitting a parametric loss function to a reconstruction of data from their plots. We find that the reported estimates are inconsistent with their first two estimation methods, fail at fitting the extracted data, and report implausibly narrow confidence intervals--intervals this narrow would require over 600,000 experiments, while they likely only ran fewer than 500. In contrast, our rederivation of the scaling law using the third app"},"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":"2404.10102","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-04-15T19:19:56Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"f3d49cda8427991e7453148ecf8b9c993d44e9b901688e60a245252908108770","abstract_canon_sha256":"06684eb544a0e1a0356850d5095d2867e78a5ed3548f08256c3aad9f85fbbaee"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:19:08.916852Z","signature_b64":"I049corJhXenaQUmHDHlYw27KOCWgFmTPmXckfKt5pUUl8TbjKM1q/2ts0873ooCkFqADu/LUWpzTs+HPGR8Cw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d293b72a6d43f1f66985f81706662f4e580fc1791a896c8ac81b8cb834489220","last_reissued_at":"2026-07-05T08:19:08.916363Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:19:08.916363Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Chinchilla Scaling: A replication attempt","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Ege Erdil, Josh You, Matthew Barnett, Tamay Besiroglu","submitted_at":"2024-04-15T19:19:56Z","abstract_excerpt":"Hoffmann et al. (2022) propose three methods for estimating a compute-optimal scaling law. We attempt to replicate their third estimation procedure, which involves fitting a parametric loss function to a reconstruction of data from their plots. We find that the reported estimates are inconsistent with their first two estimation methods, fail at fitting the extracted data, and report implausibly narrow confidence intervals--intervals this narrow would require over 600,000 experiments, while they likely only ran fewer than 500. In contrast, our rederivation of the scaling law using the third app"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.10102","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/2404.10102/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":"2404.10102","created_at":"2026-07-05T08:19:08.916421+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.10102v2","created_at":"2026-07-05T08:19:08.916421+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.10102","created_at":"2026-07-05T08:19:08.916421+00:00"},{"alias_kind":"pith_short_12","alias_value":"2KJ3OKTNIPY7","created_at":"2026-07-05T08:19:08.916421+00:00"},{"alias_kind":"pith_short_16","alias_value":"2KJ3OKTNIPY7M2MF","created_at":"2026-07-05T08:19:08.916421+00:00"},{"alias_kind":"pith_short_8","alias_value":"2KJ3OKTN","created_at":"2026-07-05T08:19:08.916421+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":20,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.25008","citing_title":"Neural Scaling Universality: If Exponents Are Fixed, Time to Understand Coefficients","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2606.24998","citing_title":"Internal Data Repetition Destroys Language Models","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01487","citing_title":"How to Allocate Your Tokens? Scaling Laws with Training Steps and Batch Size","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2606.10127","citing_title":"Data-Driven Automation","ref_index":47,"is_internal_anchor":false},{"citing_arxiv_id":"2606.01302","citing_title":"Structure and Scale in Simplicial Sequence Modelling","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31345","citing_title":"Automated High-Precision Extraction and Forensic Verification of Data-Bearing Vector Figures","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2604.24037","citing_title":"A Limit Theory of Foundation Models: A Mathematical Approach to Understanding Emergent Intelligence and Scaling Laws","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16430","citing_title":"A Theory of Training Profit-Optimal LLMs","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.24316","citing_title":"From One-Pass SGD to Data Reuse: Mini-Batch Scaling Laws in Sketched Linear Regression","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2409.04777","citing_title":"Optimization Hyper-parameter Laws for Large Language Models","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10933","citing_title":"DECO: Sparse Mixture-of-Experts with Dense-Comparable Performance on End-Side Devices","ref_index":140,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16430","citing_title":"A Theory of Training Profit-Optimal LLMs","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14200","citing_title":"How to Scale Mixture-of-Experts: From muP to the Maximally Scale-Stable Parameterization","ref_index":92,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10933","citing_title":"DECO: Sparse Mixture-of-Experts with Dense-Comparable Performance on End-Side Devices","ref_index":140,"is_internal_anchor":false},{"citing_arxiv_id":"2604.24037","citing_title":"A Limit Theory of Foundation Models: A Mathematical Approach to Understanding Emergent Intelligence and Scaling Laws","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09154","citing_title":"Predicting Large Model Test Losses with a Noisy Quadratic System","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10933","citing_title":"DECO: Sparse Mixture-of-Experts with Dense-Comparable Performance on End-Side Devices","ref_index":140,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09189","citing_title":"Practical Scaling Laws: Converting Compute into Performance in a Data-Constrained World","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2604.21106","citing_title":"How Much Is One Recurrence Worth? Iso-Depth Scaling Laws for Looped Language Models","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2505.10465","citing_title":"Superposition Yields Robust Neural Scaling","ref_index":47,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2KJ3OKTNIPY7M2MF7ALQMZRPJZ","json":"https://pith.science/pith/2KJ3OKTNIPY7M2MF7ALQMZRPJZ.json","graph_json":"https://pith.science/api/pith-number/2KJ3OKTNIPY7M2MF7ALQMZRPJZ/graph.json","events_json":"https://pith.science/api/pith-number/2KJ3OKTNIPY7M2MF7ALQMZRPJZ/events.json","paper":"https://pith.science/paper/2KJ3OKTN"},"agent_actions":{"view_html":"https://pith.science/pith/2KJ3OKTNIPY7M2MF7ALQMZRPJZ","download_json":"https://pith.science/pith/2KJ3OKTNIPY7M2MF7ALQMZRPJZ.json","view_paper":"https://pith.science/paper/2KJ3OKTN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.10102&json=true","fetch_graph":"https://pith.science/api/pith-number/2KJ3OKTNIPY7M2MF7ALQMZRPJZ/graph.json","fetch_events":"https://pith.science/api/pith-number/2KJ3OKTNIPY7M2MF7ALQMZRPJZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2KJ3OKTNIPY7M2MF7ALQMZRPJZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2KJ3OKTNIPY7M2MF7ALQMZRPJZ/action/storage_attestation","attest_author":"https://pith.science/pith/2KJ3OKTNIPY7M2MF7ALQMZRPJZ/action/author_attestation","sign_citation":"https://pith.science/pith/2KJ3OKTNIPY7M2MF7ALQMZRPJZ/action/citation_signature","submit_replication":"https://pith.science/pith/2KJ3OKTNIPY7M2MF7ALQMZRPJZ/action/replication_record"}},"created_at":"2026-07-05T08:19:08.916421+00:00","updated_at":"2026-07-05T08:19:08.916421+00:00"}