Leveraging Large Language Models to Enhance Reinforcement Learning: A Comprehensive Survey

Large language models (LLMs) can be leveraged to enhance various aspects of the reinforcement learning (RL) paradigm, including improving sample efficiency, reward function design, generalization, and natural language understanding.