Modern Large Language Models (LLMs) are often referred to as “black boxes”.