A company deploys a SageMaker model that detects topics in social media posts. They need to show how individual input features affect the model's predictions. Which SageMaker capability provides feature-level explanations for model behavior?

SageMaker Clarify is the correct choice because it offers tools for bias detection and explainability, including feature-level explanations for model predictions. It can generate SHAP (SHapley Additive exPlanations) values, which quantify the contribution of each feature to an individual prediction, directly addressing the need to understand how input features affect the model's output. SageMaker Canvas is a low-code/no-code solution for building models, not for explaining them. SageMaker Feature Store is used for creating, storing, and sharing features for machine learning models, not for model explainability. SageMaker Ground Truth is a data labeling service used to build high-quality training datasets, not for interpreting model behavior.