Breaking Entropy Bounds: Accelerating RL Training via MTP with Rejection Sampling Paper • 2606.12370 • Published Jun 10 • 21