Databricks Generative AI Engineer Associate Question 104
Select 3You are fine-tuning a generative AI model with Databricks to summarize customer support tickets. During testing, you observe that the summaries are either too detailed or too vague depending on how the prompt is structured. Which of the following prompt adjustments can help you achieve more consistent and concise summaries?
- A
Clearly specify the desired summary length, such as 'Summarize this in no more than 50 words.'
- B
Include examples of desired outputs in the prompt to guide the model's response style.
- C
Avoid using specific instructions in the prompt as they may confuse the model.
- D
Use open-ended prompts such as 'What do you think about this?' to encourage creativity.
- E
Incorporate keywords like 'concise' or 'brief' to influence the tone of the summary.
Show answer and explanation
Correct answers: A, B, E
Explanation
Prompt engineering is critical when working with generative AI models. Clear instructions, examples, and guiding keywords help models produce consistent and high-quality outputs. Ambiguous or open-ended prompts can lead to unpredictable results, especially in tasks like summarization where structure and precision are essential.
- A. Correct.
Specifying the desired summary length helps set clear expectations for the model, leading to outputs that align with your requirements.
- B. Correct.
Providing examples of desired outputs in the prompt helps the model better understand the expected format and tone, leading to more consistent results.
- C. Incorrect.
Avoiding specific instructions can lead to ambiguous outputs, as the model may not clearly understand your requirements.
- D. Incorrect.
Open-ended prompts like 'What do you think about this?' are unsuitable for tasks requiring specific outcomes, such as generating concise summaries.
- E. Correct.
Incorporating keywords like 'concise' or 'brief' provides additional guidance to the model, influencing the style and tone of the generated summaries.