Reinforcement Learning to Train Large Language Models to Explain Human Decisions arxiv.org 17 points by PaulHoule a day ago