Histograms (split by class) of the raw data, final model scores and model errors can be viewed.
 
These histograms should give a graphic demonstration of why certain variables are important and why some are not. The red and blue bars indicate the class distribution. Separation of the two colours should be more evident for the more important variables in the model.