The Dirichlet distribution produces probability vectors that can be used with discrete distributions. That is, it randomly generates a given number of values that are positive and sum to one as expected for probabilities. It has a parameter of positive, real value that controls the concentration of the probabilities. Values closer to zero mean that only a few values will be positive and receive most probability mass. The following screenshot illustrates three draws of size 10 for α = 0.1 (the dirichlet_distribution notebook contains a simulation so you can experiment with different parameter values):
Dirichlet allocation