CONFINE: Conformal Prediction for Interpretable Neural Networks