Pass@k Training for Adaptively Balancing Exploration and Exploitation of Large Reasoning Models Paper • 2508.10751 • Published Aug 14 • 28