standardize/normalize within multinom fits

[juniperlsimonis]

in some instances (especially if we include a prior on the regressors), we might need to standardize or normalize the x values within the multinomial regressions.
the option needs to be added and make sure it's carried through cleanly. (previous attempts created fitting issues when the standardization happened too early in the pipeline)