When setting bin_continuous = FALSE and use_density = TRUE an error is thrown.
This is because elements of the distribution are extracted which do not exist.
The mean and the standard deviation must be extracted in a different manner when features are scaled.
I proposed a fix in #160.
