Use Azure OpenAI in Foundry Models to Generate Content AI-102 Practice Questions (Page 2)

Questions 6–10

1. 6application · easy

   A developer has deployed a gpt-4o model in Foundry Models. The developer wants to test the model with a prompt before integrating it into an application. What should the developer use?

   • AThe Azure portal resource overview
   • BThe Azure CLI
   • CA new model deployment
   • DThe Foundry Models chat playground
   Reveal Solution Discussion Select an answer first
2. 7application · medium

   A company is building a web application that summarizes news articles. The application will use Azure OpenAI. The developer needs to integrate the model into the application code. What should the developer use to call the model?

   • AThe Foundry Models web interface
   • BThe Azure OpenAI SDK with the endpoint and API key
   • CThe Azure CLI command 'az openai chat'
   • DA direct SQL query to the Azure OpenAI database
   Reveal Solution Discussion Select an answer first
4. 8expert · hard

   A company wants to generate customer-facing content that must be factual and free of hallucinations. They have deployed gpt-4o. The team is considering using a lower temperature setting. What is the best approach to reduce hallucinations?

   • AIncrease the max_tokens parameter
   • BUse a higher temperature to encourage creativity
   • CSwitch to gpt-3.5-turbo
   • DUse a lower temperature and provide source material in the prompt
   Reveal Solution Discussion Select an answer first
5. 9application · medium

   A company wants to integrate Azure OpenAI into a customer-support chatbot that runs on Azure Functions. The chatbot must handle high concurrency and respond quickly. The team has already provisioned an Azure OpenAI resource and deployed a model. What should they do next to integrate the model into the application?

   • AUse the Azure OpenAI SDK to call the deployed model endpoint from the Azure Function.
   • BExport the model to a local container and run it in the Function.
   • CCreate a new model deployment for each Azure Function instance.
   • DUse the Azure CLI to invoke the model from the Function code.
   Reveal Solution Discussion Select an answer first
6. 10application · medium

   A support application receives user-submitted screenshots of error messages and needs to generate a plain-language explanation of the error. The team wants to use Azure OpenAI to process the images and produce text. Which model should they deploy?

   • ADeploy gpt-35-turbo and pass the screenshot as a base64-encoded image
   • BDeploy DALL-E 3, which can analyze screenshots and generate error explanations
   • CDeploy gpt-4o, which accepts image and text inputs and returns text responses
   • DDeploy text-embedding-3-large to convert screenshots into text
   Reveal Solution Discussion Select an answer first