Quiz on prompt evaluation
Time: 5–10 minutes
Rate your confidence across the prompt evaluation module. This check-in tells you which lessons are worth a second pass before you start iterating on prompt engineering techniques in the next module.
1.I can explain the difference between prompt engineering and prompt evaluation, and why both matter.
2.I can walk through the five-step eval workflow from drafting a prompt to iterating on scores.
3.I can use Claude to automatically generate an eval dataset and persist it to disk.
4.I can implement run_prompt, run_test_case, and run_eval to drive a dataset through Claude.
5.I can build a model grader that returns structured JSON with reasoning before the score.
6.I can build code graders using json.loads, ast.parse, and re.compile for syntax validation.
7.I can combine model-grader and code-grader scores into a single metric for prompt iteration.
Review any areas where your confidence is below 4 before moving on.