AI-102 Question 447
Select 3You are building a chatbot using Azure OpenAI Service. The chatbot needs to generate natural language responses to user queries. Which of the following steps must you take to submit prompts effectively and ensure high-quality responses?
- A
Design the prompt to include clear instructions and context for the task.
- B
Set a high temperature value to ensure deterministic responses.
- C
Use the most appropriate deployment model for your use case, such as 'gpt-4'.
- D
Use the Azure CLI to deploy the model and ensure the prompt is passed via the
--contextparameter. - E
Implement a feedback loop to refine prompts and improve response quality.
Show answer and explanation
Correct answers: A, C, E
Explanation
To effectively submit prompts for natural language generation, you need to design them with clarity and context, select the right model for your use case, and continuously refine prompts using feedback. These steps ensure the AI generates relevant and high-quality responses. Avoid relying on incorrect configurations (e.g., high temperature for determinism) or CLI commands not intended for prompt submission.
- A. Correct.
Designing the prompt to include clear instructions and context is crucial for the AI to understand the task and generate accurate, high-quality responses. Lack of clarity can lead to vague or incorrect outputs.
- B. Incorrect.
Setting a high temperature value does not ensure deterministic responses; instead, it increases randomness. For deterministic responses, a low temperature value should be used.
- C. Correct.
Choosing the appropriate deployment model, such as 'gpt-4', ensures that the model's capabilities align with the complexity of the task and expected quality of responses.
- D. Incorrect.
The Azure CLI is used for managing and deploying resources, but there is no specific
--contextparameter for passing prompts. Prompts are generally submitted programmatically through an API call. - E. Correct.
Implementing a feedback loop is essential for iterating on prompts and improving the quality of responses over time, especially for complex or evolving scenarios.