Great! In order to understand which features result in the anomalies, you will inspect the Summary Statistics by Cluster. It is described in “Step 3 — Get the Summary Statistics by Cluster”. Alternatively, if you want the model to be highly interpretable, you can consider the SHAP values. Click my post “Explain your model with the SHAP values” and “Explain any models with the SHAP values — use the KernelExplainer”.