Leinonen et al., 2023a ) analyzed code explanations produced by both LLMs and students, evaluating accuracy, length, and understandability. The study used three programming tasks and generated ...