Research map / Learning theory and algorithms / Kernels and regression theory
Scaling Laws from the Data Manifold Dimension
Utkarsh Sharma, Jared Kaplan
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| measure intrinsic dimension on larger student networks (ablation) | teacher/student CE · Teacher/student | intrinsic dimension: about 10% | No clear effect |
| use CE loss instead of MSE for the ID vs 1/alpha relation (ablation) | teacher/student · Teacher/student | slope of ID vs 1/alpha: 4.08 vs 3.9 | No clear effect |
| build the teacher as a product of independent sub-manifolds (T3+6) (ablation) | teacher/student · Product manifold | 4/alpha: 4/alpha 5.76 with ID 7.86 | Mixed |
| repeat a 3-feature teacher three times (T3+3+3) (ablation) | teacher/student · Product manifold | 4/alpha: 2.66 with ID 7.29 versus 3.3 for T3 alone | No clear effect |
| replace the loss with generalized |y-y*|^p (ablation) | teacher/student d about 7 · Teacher/student | alpha: alpha = 2p/d | Improved |
| measure ID from one 1024-token passage rather than many sequences (ablation) | GPT-2 117M · Language modeling | intrinsic dimension: 90 -> 7 | Hurt |
| evaluate training loss instead of test loss for CIFAR10 (ablation) | CNN · CIFAR10 | 4/alpha: 17.83 -> 14.19 | Mixed |
| use error rate instead of loss on CIFAR10 (ablation) | CNN · CIFAR10 | 4/alpha: 17.83 -> 16.23 | No clear effect |
Closest papers by meaning
- Understanding Scaling Laws with Statistical and Approximation Theory for Transformer Neural Networks on Intrinsically Low-dimensional Data (NeurIPS 2024 · similarity 0.92)
- A Dynamical Model of Neural Scaling Laws (ICML 2024 · similarity 0.91)
- Analyzing Neural Scaling Laws in Two-Layer Networks with Power-Law Data Spectra (ICLR 2025 · similarity 0.89)
- Scaling Laws in Linear Regression: Compute, Parameters, and Data (NeurIPS 2024 · similarity 0.89)
- On the Invariance and Generality of Neural Scaling Laws (NeurIPS 2026 · similarity 0.89)
- Sharp feature-learning transitions and Bayes-optimal neural scaling laws in extensive-width networks (NeurIPS 2026 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
