-
Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning
Paper • 2506.01939 • Published • 189 -
ProRL: Prolonged Reinforcement Learning Expands Reasoning Boundaries in Large Language Models
Paper • 2505.24864 • Published • 145 -
The Entropy Mechanism of Reinforcement Learning for Reasoning Language Models
Paper • 2505.22617 • Published • 130
melon
jellyisadog
AI & ML interests
None yet
Organizations
