Understanding Generative AI and Mastering the Art of Prompting Quiz

Quiz: Advanced Prompting

Pass mark: 70%. You can retry as many times as you need.

1 You will paste the AI's answer into a spreadsheet. Which instruction best protects the quality of the result? Single answer "Give me the data in a nice-looking paragraph." "Answer in a creative format of your choice." "Return a CSV table with these exact columns, no commentary, and 'null' for missing values." "Answer as quickly as possible."

2 Which are real benefits of prompt chaining (using the output of one prompt as the input of the next)? (several answers) Multiple answers possible You can check and correct each intermediate result It guarantees that the final answer is correct Each step stays simple and can be reused elsewhere You catch an error in the extraction step before it spreads to the final text

3 You ask for a persona: "Act as an experienced doctor and tell me what this symptom means." What does the persona actually change? Single answer It turns the answer into verified medical advice. It mainly changes the tone and angle of the answer, not its reliability. It gives the model access to the user's medical records.

4 You ask the model to classify customer comments by theme, and it keeps forcing odd comments into the wrong category. Which changes should help? (several answers) Multiple answers possible Add a one-line definition for each category Add an "Other / unclear" category Remove the list of categories so the model decides freely Give one or two examples of borderline comments with the right category

5 Your prompt gives good results on one test, but you suspect it is unreliable. What is the best way to evaluate it? Single answer Run it on several varied inputs and score each result against criteria defined in advance. Run the same easy input again and keep the best result. Rewrite the whole prompt at once and see if it feels better. Assume it is fine because the model answered confidently.