AI-102 Question 443
Select 3A company wants to build a natural language processing (NLP) solution using Azure OpenAI Service. They need a model that can summarize long documents and answer complex questions based on the text provided. What steps should you take to select and deploy the appropriate Azure OpenAI model for this requirement?
- A
Evaluate the GPT-4 model for its advanced text generation and comprehension capabilities.
- B
Deploy the Codex model, as it specializes in understanding and generating code.
- C
Use the Azure OpenAI playground to test various prompts and fine-tune the model behavior.
- D
Select the DALL-E model to generate visual summaries of the document.
- E
Configure the GPT model with a deployment in your Azure resource to integrate it with the application.
Show answer and explanation
Correct answers: A, C, E
Explanation
To meet the requirement of summarizing documents and answering complex questions, GPT-4 is the most suitable model due to its advanced NLP capabilities. Testing and refining prompts in the Azure OpenAI playground ensures the model behaves as needed, and deploying the model in an Azure resource is essential for application integration. Codex and DALL-E are specialized models for other domains (coding and image generation, respectively) and are not appropriate for this scenario.
- A. Correct.
GPT-4 is highly capable of handling tasks such as summarization and answering complex questions, making it a suitable choice for this scenario.
- B. Incorrect.
Codex is designed specifically for code-related tasks and is not suitable for general NLP tasks such as document summarization or question answering.
- C. Correct.
The Azure OpenAI playground allows testing and refining prompts to ensure the model behaves as expected, making it a critical step in the implementation.
- D. Incorrect.
DALL-E is a model for generating images from textual descriptions, and it is not designed for text-based NLP tasks like summarization or Q&A.
- E. Correct.
Deploying the GPT model in your Azure resource is necessary to integrate the solution with your application and make it accessible via APIs.