Exploring RAFT: The Future of AI with Retrieval-Augmented Fine-Tuning