Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning Paper • 2506.01939 • Published 5 days ago • 130
Faster Video Diffusion with Trainable Sparse Attention Paper • 2505.13389 • Published 19 days ago • 35
ReTool: Reinforcement Learning for Strategic Tool Use in LLMs Paper • 2504.11536 • Published Apr 15 • 61