Within software engineering research, Large Language Models (LLMs) are often treated as ‘black boxes’, with only their inputs and outputs being considered. In this paper, we take a machine ...