What exactly does the Bonferroni correction in the Dunn Test code do:

Code:
 dunntest rating, by(outcome) ma(bonferroni)
With the regular two-sample Wilcoxon rank-sum (Mann-Whitney) test, you need to multiple the p-value by the number of groups to get the Bonferroni corrected p-value. Does the Dunn test already do this for us?