Answer: Source Attribution in Knowledge Bases

**Correct Answer: B (Source Attribution in Knowledge Bases)** **Explanation:** Source Attribution in Knowledge Bases is specifically designed to provide citations and references to the exact source documents used to generate AI responses. This feature is crucial for legal teams and other professionals who need to verify the accuracy and provenance of information provided by AI systems. **Why other options are incorrect:** - **A) Guardrails for Bedrock**: While Guardrails help ensure AI responses are safe, appropriate, and aligned with policies, they don't specifically provide source citations. - **C) Model Evaluation**: This involves assessing model performance, accuracy, and bias, but doesn't provide source attribution for individual responses. - **D) Reinforcement Learning**: This is a machine learning training technique where models learn through trial and error with rewards/penalties, not a feature for source citation. **Key Benefits of Source Attribution:** 1. **Transparency**: Users can see exactly which documents were used to generate the answer 2. **Verifiability**: Legal teams can verify information against original sources 3. **Trust**: Builds confidence in AI-generated responses by showing their provenance 4. **Compliance**: Helps meet regulatory requirements for traceability and accountability

Author: Jin H