the distribution plot is the only one that does not set a x-label Perhaps a parameter would be nice, e.g. ``` audplot.distribution( truth, prediction, x_label = ' likability' ) ```