Scaling Laws papers, 2023–2026

Key findings

• Scaling Laws went from 60 papers in 2023 to 248 in 2026, roughly 4.1x in three years.
• NeurIPS publishes the most Scaling Laws work: 221 papers from 2023 to 2026, 37% of the four-venue total.
• The fastest-growing overlap is with Time Series: papers tagged with both rose from 1 in 2023 to 24 in 2026.
• Among the 5 topics under Optimization and Scaling, it ranks #5 by 2026 paper count; the fastest grower in the group, Low-Rank Adaptation, is up 3.0x versus 4.1x here.

Representative papers

Spread across years and venues so every edition shows up

• NeurIPS 2026A Unified Merge Calculus for Learning-Rate Scaling on Neural Computation Graphs
• ICML 2026A Capacity-Based Rationale for Multi-Head Attention
• ICLR 2026$\mathbf{Li_2}$: A Framework on Dynamics of Feature Emergence and Delayed Generalization
• CVPR 2026A Unified Framework for Knowledge Transfer in Bidirectional Model Scaling
• NeurIPS 2025Audits Under Resource, Data, and Access Constraints: Scaling Laws For Less Discriminatory Alternatives
• ICML 2025A Simple Model of Inference Scaling Laws
• ICLR 2025(Mis)Fitting Scaling Laws: A Survey of Scaling Law Fitting Techniques in Deep Learning
• CVPR 2025D^3: Scaling Up Deepfake Detection by Learning from Discrepancy
• NeurIPS 2024An exactly solvable model for emergence and scaling laws in the multitask sparse parity problem
• ICML 2024A Dynamical Model of Neural Scaling Laws
• ICLR 2024Principled Architecture-aware Scaling of Hyperparameters
• CVPR 2024From Activation to Initialization: Scaling Insights for Optimizing Neural Fields
• NeurIPS 2023Abide by the law and follow the flow: conservation laws for gradient flows
• ICML 2023Dissecting the Effects of SGD Noise in Distinct Regimes of Deep Learning
• ICLR 2023Distributionally Robust Post-hoc Classifiers under Prior Shifts
• CVPR 2023LargeKernel3D: Scaling Up Kernels in 3D Sparse CNNs
• NeurIPS 2026ADIS-Law: Unified Scaling Laws for Annealing-Phase Domain Injection in Large Language Models
• ICML 2026Balancing Learning Rates Across Layers: Exact Two-Step Dynamics and Optimal Scaling in Linear Neural Networks
• ICLR 2026A universal compression theory: Lottery ticket hypothesis and superpolynomial scaling laws
• CVPR 2026MatAnyone 2 Scaling Video Matting via a Learned Quality Evaluator

All Scaling Laws papers

605 papers, page 7 of 13

301. ICLR 2025Scaling Optimal LR Across Token Horizons
302. ICLR 2025Scaling Stick-Breaking Attention: An Efficient Implementation and In-depth Study
303. ICLR 2025Scaling up Masked Diffusion Models on Text
304. ICLR 2025Scaling Wearable Foundation Models
305. ICLR 2025SimBa: Simplicity Bias for Scaling Up Parameters in Deep Reinforcement Learning
306. ICLR 2025Strong Model Collapse
307. ICLR 2025Surprising Effectiveness of pretraining Ternary Language Model at Scale
308. ICLR 2025The Journey Matters: Average Parameter Count over Pre-training Unifies Sparse and Dense Scaling Laws
309. ICLR 2025TokenFormer: Rethinking Transformer Scaling with Tokenized Model Parameters
310. ICLR 2025Towards Neural Scaling Laws for Time Series Foundation Models
311. ICLR 2025Transformers Struggle to Learn to Search
312. ICLR 2025U-shaped and Inverted-U Scaling behind Emergent Abilities of Large Language Models
313. ICLR 2025Ultra-Sparse Memory Network
314. ICML 2025A Forget-and-Grow Strategy for Deep Reinforcement Learning Scaling in Continuous Control
315. ICML 2025A Hitchhiker's Guide to Scaling Law Estimation
316. ICML 2025A Simple Model of Inference Scaling Laws
317. ICML 2025Bayesian Neural Scaling Law Extrapolation with Prior-Data Fitted Networks
318. ICML 2025Contextures: Representations from Contexts
319. ICML 2025DataDecide: How to Predict Best Pretraining Data with Small Experiments
320. ICML 2025Determinant Estimation under Memory Constraints and Neural Scaling Laws
321. ICML 2025Distillation Scaling Laws
322. ICML 2025Does Data Scaling Lead to Visual Compositional Generalization?
323. ICML 2025Dynamic Sparse Training of Diagonally Sparse Networks
324. ICML 2025Equivariant Polynomial Functional Networks
325. ICML 2025Haste Makes Waste: A Simple Approach for Scaling Graph Neural Networks
326. ICML 2025How Do Large Language Monkeys Get Their Power (Laws)?
327. ICML 2025How Much Can We Forget about Data Contamination?
328. ICML 2025Improving Diversity in Language Models: When Temperature Fails, Change the Loss
329. ICML 2025Improving the Scaling Laws of Synthetic Data with Deliberate Practice
330. ICML 2025Joint MoE Scaling Laws: Mixture of Experts Can Be Memory Efficient
331. ICML 2025Learning curves theory for hierarchically compositional data with power-law distributed features
332. ICML 2025Learning to Keep a Promise: Scaling Language Model Decoding Parallelism with Learned Asynchronous Decoding
333. ICML 2025LensLLM: Unveiling Fine-Tuning Dynamics for LLM Selection
334. ICML 2025LLMs on the Line: Data Determines Loss-to-Loss Scaling Laws
335. ICML 2025Memory Layers at Scale
336. ICML 2025Model Steering: Learning with a Reference Model Improves Generalization Bounds and Scaling Laws
337. ICML 2025Network Sparsity Unlocks the Scaling Potential of Deep Reinforcement Learning
338. ICML 2025No Free Lunch from Random Feature Ensembles: Scaling Laws and Near-Optimality Conditions
339. ICML 2025On Temperature Scaling and Conformal Prediction of Deep Classifiers
340. ICML 2025Over-Tokenized Transformer: Vocabulary is Generally Worth Scaling
341. ICML 2025OWLS: Scaling Laws for Multilingual Speech Recognition and Translation Models
342. ICML 2025Parameters vs FLOPs: Scaling Laws for Optimal Sparsity for Mixture-of-Experts Language Models
343. ICML 2025Parametric Scaling Law of Tuning Bias in Conformal Prediction
344. ICML 2025Penalizing Infeasible Actions and Reward Scaling in Reinforcement Learning with Offline Data
345. ICML 2025Position: AI Scaling: From Up to Down and Out
346. ICML 2025Position: Enough of Scaling LLMs! Lets Focus on Downscaling
347. ICML 2025Position: Scaling LLM Agents Requires Asymptotic Analysis with LLM Primitives
348. ICML 2025Residual Matrix Transformers: Scaling the Size of the Residual Stream
349. ICML 2025Scaling Collapse Reveals Universal Dynamics in Compute-Optimally Trained Neural Networks
350. ICML 2025Scaling Inference-Efficient Language Models

FAQ

How many Scaling Laws papers were published at NeurIPS, ICLR, ICML and CVPR from 2023 to 2026?

605 in total: 60 in 2023, 96 in 2024, 201 in 2025, 248 in 2026.

How many Scaling Laws papers are at NeurIPS 2026?

NeurIPS 2026 has 82 Scaling Laws papers, versus 80 the year before.

How many Scaling Laws papers are at ICLR 2026?

ICLR 2026 has 73 Scaling Laws papers, versus 48 the year before.

How many Scaling Laws papers are at ICML 2026?

ICML 2026 has 78 Scaling Laws papers, versus 56 the year before.

How many Scaling Laws papers are at CVPR 2026?

CVPR 2026 has 15 Scaling Laws papers, versus 17 the year before.

Which conference publishes the most Scaling Laws papers?

NeurIPS, with 221 papers from 2023 to 2026.