In the first experiment, I tested whether LLMs could act as code reviewers, asking them to define a rubric for high-quality code and then critique a given implementation based on their own criteria.