問題文
Which two activities belong to the Protect stage of the recommended responsible AI development steps for Microsoft Foundry? (Select two.)
選択肢
- Configuring Guardrails so that unsafe agent actions are prevented at runtime
- Setting up continuous monitoring to detect anomalies in production behavior
- Publishing a transparency note that describes the intended uses and limitations
- Applying content filters to block harmful model outputs before they reach users
- Testing the agent with adversarial prompts to find jailbreak vulnerabilities