Official implementation for the paper "Toward Scientific Reasoning in LLMs: Training from Expert Discussions via Reinforcement Learning" - View it on GitHub