缩放定律论文趋势 2023–2026

数据/模型缩放

统计范围：NeurIPS、ICLR、ICML、CVPR 四个会议，2023–2026 年，共 605 篇被归入该主题的论文。

主要结论

每年论文数

各会议论文数

按年份和会议统计

缩放定律论文数，按年份和会议
年份NeurIPSICLRICMLCVPR合计
2023211417860
20243815331096
202580485617201
202682737815248

代表论文

按年份和会议均匀挑选，方便快速看到每一届的样貌

全部缩放定律论文

年份全部2023202420252026
会议全部NeurIPSICLRICMLCVPR

共 605 篇，第 7 / 13 页

  1. ICLR 2025Scaling Optimal LR Across Token Horizons
  2. ICLR 2025Scaling Stick-Breaking Attention: An Efficient Implementation and In-depth Study
  3. ICLR 2025Scaling up Masked Diffusion Models on Text
  4. ICLR 2025Scaling Wearable Foundation Models
  5. ICLR 2025SimBa: Simplicity Bias for Scaling Up Parameters in Deep Reinforcement Learning
  6. ICLR 2025Strong Model Collapse
  7. ICLR 2025Surprising Effectiveness of pretraining Ternary Language Model at Scale
  8. ICLR 2025The Journey Matters: Average Parameter Count over Pre-training Unifies Sparse and Dense Scaling Laws
  9. ICLR 2025TokenFormer: Rethinking Transformer Scaling with Tokenized Model Parameters
  10. ICLR 2025Towards Neural Scaling Laws for Time Series Foundation Models
  11. ICLR 2025Transformers Struggle to Learn to Search
  12. ICLR 2025U-shaped and Inverted-U Scaling behind Emergent Abilities of Large Language Models
  13. ICLR 2025Ultra-Sparse Memory Network
  14. ICML 2025A Forget-and-Grow Strategy for Deep Reinforcement Learning Scaling in Continuous Control
  15. ICML 2025A Hitchhiker's Guide to Scaling Law Estimation
  16. ICML 2025A Simple Model of Inference Scaling Laws
  17. ICML 2025Bayesian Neural Scaling Law Extrapolation with Prior-Data Fitted Networks
  18. ICML 2025Contextures: Representations from Contexts
  19. ICML 2025DataDecide: How to Predict Best Pretraining Data with Small Experiments
  20. ICML 2025Determinant Estimation under Memory Constraints and Neural Scaling Laws
  21. ICML 2025Distillation Scaling Laws
  22. ICML 2025Does Data Scaling Lead to Visual Compositional Generalization?
  23. ICML 2025Dynamic Sparse Training of Diagonally Sparse Networks
  24. ICML 2025Equivariant Polynomial Functional Networks
  25. ICML 2025Haste Makes Waste: A Simple Approach for Scaling Graph Neural Networks
  26. ICML 2025How Do Large Language Monkeys Get Their Power (Laws)?
  27. ICML 2025How Much Can We Forget about Data Contamination?
  28. ICML 2025Improving Diversity in Language Models: When Temperature Fails, Change the Loss
  29. ICML 2025Improving the Scaling Laws of Synthetic Data with Deliberate Practice
  30. ICML 2025Joint MoE Scaling Laws: Mixture of Experts Can Be Memory Efficient
  31. ICML 2025Learning curves theory for hierarchically compositional data with power-law distributed features
  32. ICML 2025Learning to Keep a Promise: Scaling Language Model Decoding Parallelism with Learned Asynchronous Decoding
  33. ICML 2025LensLLM: Unveiling Fine-Tuning Dynamics for LLM Selection
  34. ICML 2025LLMs on the Line: Data Determines Loss-to-Loss Scaling Laws
  35. ICML 2025Memory Layers at Scale
  36. ICML 2025Model Steering: Learning with a Reference Model Improves Generalization Bounds and Scaling Laws
  37. ICML 2025Network Sparsity Unlocks the Scaling Potential of Deep Reinforcement Learning
  38. ICML 2025No Free Lunch from Random Feature Ensembles: Scaling Laws and Near-Optimality Conditions
  39. ICML 2025On Temperature Scaling and Conformal Prediction of Deep Classifiers
  40. ICML 2025Over-Tokenized Transformer: Vocabulary is Generally Worth Scaling
  41. ICML 2025OWLS: Scaling Laws for Multilingual Speech Recognition and Translation Models
  42. ICML 2025Parameters vs FLOPs: Scaling Laws for Optimal Sparsity for Mixture-of-Experts Language Models
  43. ICML 2025Parametric Scaling Law of Tuning Bias in Conformal Prediction
  44. ICML 2025Penalizing Infeasible Actions and Reward Scaling in Reinforcement Learning with Offline Data
  45. ICML 2025Position: AI Scaling: From Up to Down and Out
  46. ICML 2025Position: Enough of Scaling LLMs! Lets Focus on Downscaling
  47. ICML 2025Position: Scaling LLM Agents Requires Asymptotic Analysis with LLM Primitives
  48. ICML 2025Residual Matrix Transformers: Scaling the Size of the Residual Stream
  49. ICML 2025Scaling Collapse Reveals Universal Dynamics in Compute-Optimally Trained Neural Networks
  50. ICML 2025Scaling Inference-Efficient Language Models

在 Claude、Cursor 里直接问这个趋势

接上 Paperlist MCP 后，写综述、挑方向或准备投稿时，不用离开对话框就能拿到这页的数字：每年各会议的论文数、增长倍数、涨得最快的子方向和代表论文，AI 会按你的问题整理。

可以直接复制这句话问 AI

用 paperlist 查缩放定律在 NeurIPS 2023–2026 的论文趋势，哪一年涨得最快？

AI 会调用的工具：get_topic_stats {"topic": "scaling-laws"}

两分钟接入 Paperlist MCP（Claude Code、Cursor、Codex、ChatGPT）

常见问题

2023–2026 年四大会议一共有多少篇缩放定律论文？
共 605 篇，按年份分别是 2023 年 60 篇、2024 年 96 篇、2025 年 201 篇、2026 年 248 篇。
NeurIPS 2026 有多少篇缩放定律论文？
NeurIPS 2026 有 82 篇缩放定律论文，上一届是 80 篇。
ICLR 2026 有多少篇缩放定律论文？
ICLR 2026 有 73 篇缩放定律论文，上一届是 48 篇。
ICML 2026 有多少篇缩放定律论文？
ICML 2026 有 78 篇缩放定律论文，上一届是 56 篇。
CVPR 2026 有多少篇缩放定律论文？
CVPR 2026 有 15 篇缩放定律论文，上一届是 17 篇。
哪个会议的缩放定律论文最多？
NeurIPS，2023–2026 年共 221 篇。

「优化与缩放」下的其他主题