Taming LLMs by Scaling Learning Rates with Gradient Grouping Paper • 2506.01049 • Published Jun 1 • 37
