Basics of Prompt Engineering AIE Practice Questions (Page 4)

Questions 16–20

1. 16application · medium

   A data scientist is using an AI to generate Python code for data visualization. The first output used a library that is not installed in the environment. The data scientist wants to improve the prompt. What should they do?

   • AInstall the library that the AI suggested.
   • BAdd a constraint to the prompt: 'Use only libraries from this list: matplotlib, pandas, numpy.'
   • CAsk the AI to explain why it chose that library.
   • DUse a different AI model that is more familiar with the environment.
   Reveal Solution Discussion Select an answer first
2. 17expert · hard

   A multilingual company wants an AI to translate customer feedback into English. The AI sometimes produces translations that are too literal and miss cultural nuances. The team wants to improve translation quality without providing a glossary for every product. Which prompt component is most valuable to add?

   • AA few-shot example of a literal translation.
   • BA list of common idioms in the source language.
   • CA request for a word-for-word translation.
   • DAn instruction to 'translate meaning, not words, and adapt to English cultural context'.
   Reveal Solution Discussion Select an answer first
4. 18application · medium

   A project manager is using an AI assistant to summarize meeting notes. The first summary was too verbose and missed action items. The manager wants to evaluate and improve the prompt. Which action best reflects evaluating prompt effectiveness?

   • AAsk the AI to rate its own summary on a scale of 1 to 10.
   • BChange the prompt randomly to see if the output improves.
   • CCompare the summary against the original notes to check if all action items are captured and the length is appropriate.
   • DUse the same prompt again and hope the next summary is better.
   Reveal Solution Discussion Select an answer first
5. 19foundation · easy

   What is a zero-shot prompt?

   • AA prompt that provides no examples and relies on the model's pre-trained knowledge to perform the task
   • BA prompt that uses a special token to reset the model's memory
   • CA prompt that includes one example to guide the model's response
   • DA prompt that instructs the model to generate zero responses
   Reveal Solution Discussion Select an answer first
6. 20expert · hard

   A team is using an AI to generate product descriptions for an e-commerce site. The first prompt produced descriptions that were too long and did not include the brand voice. The team has a limited budget for API calls and wants to minimize the number of iterations. Which approach best balances effectiveness and cost?

   • AUse a short prompt and iterate multiple times based on each response.
   • BUse a zero-shot prompt and accept the first output as-is.
   • CUse a chain-of-thought prompt to have the model explain its reasoning before writing.
   • DStart with a detailed prompt that includes the brand voice guidelines, a length limit, and a few examples of desired descriptions.
   Reveal Solution Discussion Select an answer first