3 ms·Eureka: Human-Level Reward Design via Coding Large Language Models2 points by famouswaffles 3y ago