Default priors across families

[athowes]

In #155 we added messaging about the default priors being used. This messaging could be improved by including information about what the priors imply about an interpretable parameter e.g. on the natural scale. This is difficult because we don't currently have the family information in epidist_prior which e.g. would be required to generate natural parameters (#169). Additionally, we may want to be setting different priors for different families. The parameters we need to use may not be as similar as we once thought (#131).

[athowes]

We now have epidist_family_prior as of #242.