Explanation:

Source Attribution in Knowledge Bases is the correct feature that enables AI answers to cite the exact document used as a source. This feature is particularly important for legal teams and other compliance-focused organizations that require traceability and verifiability of information sources.

Why other options are incorrect:

• A) Guardrails for Bedrock: This feature focuses on content safety, filtering inappropriate content, and implementing safety controls, not specifically on source attribution.
• C) Model Evaluation: This involves assessing model performance, accuracy, and metrics, not source citation capabilities.
• D) Reinforcement Learning: This is a machine learning training methodology where models learn through trial and error with rewards, unrelated to source attribution.

Key Benefits of Source Attribution:

1. Transparency: Shows exactly which documents were used to generate answers
2. Verifiability: Allows users to check the original source material
3. Compliance: Meets legal and regulatory requirements for source documentation
4. Trust: Builds confidence in AI-generated responses by providing traceability

This feature is essential for applications where audit trails, compliance, and source verification are critical requirements.