On the Limitations of Large Language Models (LLMs): False Attribution