Robust Layerwise Scaling Rules by Proper Weight Decay Tuning Paper • 2510.15262 • Published Oct 17 • 5
On the Design of KL-Regularized Policy Gradient Algorithms for LLM Reasoning Paper • 2505.17508 • Published May 23 • 7
MARS: Unleashing the Power of Variance Reduction for Training Large Models Paper • 2411.10438 • Published Nov 15, 2024 • 13