NCA-GENM Question 192
Select 3You are using a generative AI multimodal model to create descriptive captions for images. The initial prompt, 'Describe this image,' generates vague outputs that lack detail. How can you modify the prompt to improve the specificity and quality of the model's output?
- A
Add specific instructions like 'Focus on colors, objects, and emotions in the image.'
- B
Use a shorter prompt to reduce ambiguity, such as 'Caption this.'
- C
Provide context by including an example prompt and its desired output.
- D
Incorporate constraints like 'Generate captions no longer than 50 words.'
- E
Repeat the original prompt multiple times to emphasize its importance.
Show answer and explanation
Correct answers: A, C, D
Explanation
Effective prompt engineering involves adding clarity, specificity, and context to guide generative AI models toward producing desired outputs. By including detailed instructions, examples, and constraints, users can better influence the model's behavior. Simply shortening or repeating the prompt does not address the underlying issue of vague or insufficient guidance.
- A. Correct.
Adding specific instructions helps guide the model to focus on particular aspects of the image, improving the relevance and detail of the output.
- B. Incorrect.
Using a shorter prompt like 'Caption this' is likely to introduce more ambiguity, as it provides even less guidance to the model.
- C. Correct.
Providing context through examples helps the model understand the expected style and detail of the output, leading to better results.
- D. Correct.
Incorporating constraints like word limits ensures the output adheres to specific requirements, making it more aligned with the user's needs.
- E. Incorrect.
Repeating the original prompt does not add any new information or guidance for the model and is unlikely to influence the output quality.