Uncovering Latent Memories: Assessing Data Leakage and Memorization Patterns in Large Language Models