Basics of Prompt Engineering AIE Practice Questions (Page 7)

Questions 31–35

1. 31expert · hard

   A medical research team is using an AI to classify research abstracts into categories. They have a small set of labeled abstracts (20 per category) and need high accuracy. However, the AI model has a token limit that restricts the prompt length. Which approach best balances accuracy and token constraints?

   • AUse chain-of-thought prompting, asking the model to reason about each classification.
   • BUse few-shot prompting with a few carefully selected examples per category, keeping the prompt within the token limit.
   • CUse all 20 examples per category in the prompt, ignoring the token limit.
   • DUse zero-shot prompting with a detailed description of each category.
   Reveal Solution Discussion Select an answer first
2. 32application · medium

   A non-technical manager wants to use an AI to generate a project status report. The manager is unsure how to phrase the request. Which approach best demonstrates the role of prompt engineering?

   • AUse a vague prompt like 'Write a report'.
   • BUse a single-word command like 'Report'.
   • CProvide the project data and specify the audience, format, and key sections.
   • DAsk the model to 'do its best' without details.
   Reveal Solution Discussion Select an answer first
4. 33expert · hard

   A global company is using an AI to translate customer support emails into multiple languages. The AI sometimes produces translations that are too formal or too casual. The company wants consistent tone across all languages. Which prompt component is most critical to add?

   • AAn instruction to translate accurately without specifying tone.
   • BA request for the model to explain its translation choices.
   • CA few examples of the desired tone in each target language.
   • DA zero-shot prompt with a general statement like 'Be professional'.
   Reveal Solution Discussion Select an answer first
5. 34application · medium

   A small business owner wants to use an AI chatbot to answer customer questions about store hours and return policies. They notice the chatbot sometimes gives outdated information. What is the most effective way to apply prompt engineering to improve accuracy?

   • AAsk the chatbot to be more confident in its answers.
   • BInclude the current store hours and return policy in the system prompt.
   • CUse a longer prompt with more general knowledge about retail.
   • DRetrain the model on all customer queries.
   Reveal Solution Discussion Select an answer first
6. 35application · medium

   A non-profit organization wants to use an AI to draft grant proposals. They have a detailed template and past successful proposals. What is the most effective way to apply prompt engineering?

   • AAsk the AI to write a grant proposal from scratch without any input.
   • BUse the AI's default settings and hope for the best.
   • CProvide the AI with the template, a past proposal as context, and a clear instruction to follow the template structure.
   • DGive the AI a short prompt like 'Write a grant proposal'.
   Reveal Solution Discussion Select an answer first