Assisting Reinforcement Learning Agent Training Using Large Language Models