Predictions into Human-Interpretable Logical Rules