Reinforced Self-play Reasoning with Zero Data 论文解读

Voyager: An Open-Ended Embodied Agent with Large Language Models