Design a Prompt That Elicits a Specifically Formatted Response GENERATIVE-AI-ENGINEER-ASSOCIATE Practice Questions (Page 2)

Questions 6–10

1. 6expert · hard

   A team is building a prompt to extract the 'total' from a receipt. The receipt may contain multiple totals (e.g., subtotal, tax, total). The current prompt is: 'Extract the total from this receipt.' The model sometimes returns the subtotal instead of the final total. The team has tried adding 'Extract the final total' but the issue persists. What is the most effective way to disambiguate which total to extract?

   • AAdd 'Extract the largest number on the receipt.'
   • BAdd 'Extract the total that appears at the bottom of the receipt.'
   • CAdd 'Extract the total amount that the customer pays.'
   • DAdd a few-shot example showing a receipt with subtotal, tax, and total, and the corresponding output with only the final total.
   Reveal Solution Discussion Select an answer first
2. 7application · medium

   A support team is building a chatbot that classifies customer feedback into one of three categories: 'bug', 'feature_request', or 'other'. The downstream system only accepts these exact lowercase strings. The team's current prompt asks the model to 'classify the feedback' and the model sometimes returns labels like 'Bug', 'Feature Request', or 'Other'. Which prompt change best enforces the required output?

   • AAdd 'Classify the feedback carefully and accurately.'
   • BAdd 'Return the classification in a JSON object.'
   • CAdd 'Respond with one of these exact values: bug, feature_request, other. Do not use any other words.'
   • DAdd 'Use the category names exactly as they appear in the training data.'
   Reveal Solution Discussion Select an answer first
4. 8expert · hard

   A team is building a system that extracts structured data from legal contracts. The output must be a JSON object with fields `party_name` (string), `effective_date` (string in YYYY-MM-DD format), and `obligations` (array of strings). The team has a working prompt that produces the correct JSON structure, but the model sometimes returns `effective_date` as 'March 15, 2024' instead of '2024-03-15'. The team has already tried adding 'Use YYYY-MM-DD format for the date' to the user prompt, but the issue persists. Which additional technique is most likely to resolve the date format inconsistency?

   • ARemove the date field from the JSON output and let the downstream system parse it.
   • BIncrease the temperature to 0.7 to make the model more creative with dates.
   • CAdd 'Be very careful with the date format' to the system prompt.
   • DAdd a few-shot example showing a contract snippet with the date 'March 15, 2024' and the corresponding JSON output with 'effective_date': '2024-03-15'.
   Reveal Solution Discussion Select an answer first
5. 9foundation · easy

   Which prompt instruction is most likely to produce a markdown-formatted response?

   • AWrite a summary in a professional tone.
   • BWrite a summary as a JSON object.
   • CWrite a summary.
   • DWrite a summary using markdown with a heading and bullet points.
   Reveal Solution Discussion Select an answer first
6. 10foundation · easy

   A prompt engineer tests a prompt and finds that the model outputs markdown instead of the requested JSON. What is the most effective way to iterate on the prompt?

   • AIncrease the model's temperature to 1.5.
   • BIgnore the output and test again.
   • CRemove the word 'JSON' from the prompt.
   • DAdd a few-shot example of a JSON output in the prompt.
   Reveal Solution Discussion Select an answer first