Toward Scientific Reasoning in LLMs: Training from Expert Discussions via Reinforcement Learning Paper • 2505.19501 • Published May 26, 2025 • 1
