Infinite Sampling: Efficient and Stable Grouped RL Training for Large Language Models

Open in new window