An AI practitioner needs to evaluate models and provide explanations of predictions to customers and stakeholders. Which AWS feature or service provides explanation capabilities for model predictions?

Amazon SageMaker Clarify provides tools to help AI practitioners understand and explain model predictions, detect bias, and improve model transparency. It generates various explainability reports, such as SHAP (SHapley Additive exPlanations) values, to show how individual features contribute to a model's output, which is crucial for explaining predictions to customers and stakeholders. Amazon QuickSight is a business intelligence service for data visualization, not model explanation. Amazon Comprehend is a natural language processing service for text analysis, not general model explainability. AWS Trusted Advisor provides recommendations to optimize AWS resources, security, and cost, unrelated to AI model explanations.