Large Language Model Integration with Reinforcement Learning to Augment Decision-Making in Autonomous Cyber Operations