How to retrieve GLM hidden parameters

Hi, 
in a project I'm using the GLM operator  (based on H2O algorithm). In particular, I would like to train a Negative Binomial model and in a second step extract the theta parameter (see here for more details). I've already seen that this parameter is not reported in the result. Do you have any idea to extract it? I was thinking if it could be useful to use the Execute Script with some Java code.

thanks,

Raul