Improving Data and Reward Design for Scientific Reasoning in Large Language Models Paper • 2602.08321 • Published Feb 9 • 43
