feat: add model interpretability with SHAP explanations and uncertainty quantification dc99b5f rb757 commited on Dec 18, 2025