Right now, OPN/CPN networks can run into problems for some of the models that we can simulate via `ssm-simulators`. Two steps: - [ ] we need to take stock of any such models that may create problems (close via #27 ) - [ ] we need to do the actual work to generalize out networks to deal with n-choice scenarios (first problem: `bcelogit` loss)