Which Azure service acts as a "Global Load Balancer" to route AI traffic to the healthiest and closest regional OpenAI endpoint?
check_circle
Correct AnswerD
Azure Front Door (often paired with APIM) provides global routing and failover capabilities for high-availability AI applications.
Question 62
You are evaluating an agent's response for "Relevance". What specifically are you comparing?
check_circle
Correct AnswerD
Relevance measures how well the model's response addresses the user's intent as expressed in the initial prompt.
Question 63
You want to use a "Small Language Model" (SLM) to perform simple sentiment classification on the edge. Which model should you choose?
check_circle
Correct AnswerA
Phi-3 is a Microsoft-developed SLM optimized for high performance with a smaller footprint, ideal for simpler tasks or edge deployment.
Question 64
You are evaluating an agent's response for "Fluency". What are you specifically checking?
check_circle
Correct AnswerA
Fluency measures the linguistic quality of the output, checking if it sounds natural and follows grammar rules.
Question 65
You are configuring a Content Filter. You want to block content that contains "Self-Harm" but only if it is of "High" severity. What threshold should you set?
check_circle
Correct AnswerC
Setting the filter to "High" means only content with a High severity rating is blocked; Low and Medium are allowed.
Question 66
Which Azure AI service provides the "Personalized Voice" feature, allowing an agent to speak in a voice cloned from a 1-minute sample?
check_circle
Correct AnswerD
The Personal Voice feature in Azure AI Speech creates high-fidelity synthetic voices from small amounts of training data.
Question 67
You are evaluating an agent that is "Too Verbose." Which parameter should you decrease to encourage shorter answers?
check_circle
Correct AnswerD
Setting a lower max_tokens limit forces the model to be more concise.
Question 68
In the "Azure AI Agent Service", what is the primary benefit of using "Code Interpreter" for data analysis over a standard LLM prompt?
check_circle
Correct AnswerD
LLMs are prone to calculation errors; Code Interpreter ensures accuracy by executing actual code to perform the math.
Question 69
What is the primary role of a "Service Connection" in the context of Azure AI Foundry?
check_circle
Correct AnswerA
Service Connections (or Hub Connections) centralize the management of API keys and endpoints for use across multiple projects.
Question 70
Which metric measures how "Helpful" and "Safe" the model's response is according to human-defined guidelines?
check_circle
Correct AnswerA
Manual Evaluation in AI Foundry allows human reviewers to provide qualitative scores based on complex criteria like helpfulness and safety.
