Reinforcement Learning to Train Large Language Models to Explain Human Decisions arxiv.org 24 points by PaulHoule 3 days ago