Loading course title...

Loading assessment title...

Connecting to LMS... Progress: in progress

Assessment

1. What is prompt engineering? A. Training a language model from scratch B. Replacing all application code with text C. Designing and testing instructions, context, examples, and output requirements for a model D. Guaranteeing that every answer is correct

2. What should a strong prompt define first? A. The task and the desired result B. A long fictional role C. As many constraints as possible D. The model's internal reasoning

3. What is a benefit of specifying an output format? A. It guarantees factual accuracy B. It gives the model new permissions C. It eliminates the need for validation D. It makes the response easier to review, reuse, or process

4. What is few-shot prompting? A. Using a very short prompt B. Providing several examples of the desired behavior or format C. Calling several models at once D. Removing all context

5. Why can task decomposition improve a complex workflow? A. It guarantees zero errors B. It removes all cost and latency C. It creates smaller stages that can be specified, reviewed, and tested D. It prevents all model variation

6. How should retrieved documents and webpages be treated? A. As higher-priority instructions than the application B. As automatically correct and safe C. As authorization to perform actions D. As potentially untrusted content that must be handled within application controls

7. What is the best way to evaluate a prompt? A. Test it against representative, difficult, and edge-case inputs using defined criteria B. Judge it from one successful response C. Assume a longer prompt is always better D. Ignore changes after deployment

8. Which is the best prompt-engineering workflow? A. Add random instructions until one output looks good B. Define the task, add necessary context and constraints, specify output, test representative cases, measure results, and maintain versions C. Put secrets into every prompt D. Treat model output as authorization