AI-102 exam dumps

AI-102 practice question 471 of 493

Designing and Implementing a Microsoft Azure AI Solution. Professional level, Microsoft. Free question with the correct answer and a full explanation.

AI-102 Question 471

Select 3

You are building a generative AI solution using Azure OpenAI Service to generate product descriptions for an e-commerce platform. The current model generates overly verbose descriptions that do not align with the platform's style guidelines. What steps can you take to optimize the model's output?

  1. A

    Adjust the temperature parameter to control the randomness of the output.

  2. B

    Fine-tune the model with a custom dataset containing examples of concise product descriptions.

  3. C

    Increase the token limit to allow the model to generate longer outputs.

  4. D

    Use prompt engineering to include specific instructions for generating concise text.

  5. E

    Switch to a larger pre-trained model for more detailed outputs.

Show answer and explanation

Correct answers: A, B, D

Explanation

To optimize generative AI outputs in Azure OpenAI Service, you can adjust parameters like temperature to control randomness, fine-tune the model to align with specific style guidelines, and use prompt engineering to provide clear instructions. These approaches help address the issue of verbosity while ensuring the output meets the desired requirements. Increasing token limits or using larger models would not resolve the problem and may even worsen it.

  • A. Correct.

    Adjusting the temperature parameter can help control the randomness of the output, which can contribute to more concise and focused text generation.

  • B. Correct.

    Fine-tuning the model with a custom dataset ensures the model learns from examples that align with the desired style and tone, making the generated descriptions more concise and appropriate.

  • C. Incorrect.

    Increasing the token limit would likely exacerbate the issue of verbosity rather than optimize the output for conciseness.

  • D. Correct.

    Using prompt engineering to include clear instructions for generating concise text can effectively guide the model toward producing shorter, more relevant outputs.

  • E. Incorrect.

    Switching to a larger pre-trained model may result in more detailed outputs, which contradicts the goal of generating concise descriptions.

Timed practice exam

Take a AI-102 practice test under exam conditions

40 questions in 120 minutes, drawn from this bank, with a score report and a per-question review when you finish.

Start timed exam