Default priors for saturation and adstock functions

[wd60622]

Follow #632, the default priors for the new saturations might want to be looked at again. For instance, do the prior curves cover approximately the same area? Or are they all widely different?

For instance, after #722 this could be easy to compare.

saturations = [
    MichaelisMentenSaturation(), 
    LogisticSaturation(), 
    ..., 
]
ax = plt.gca()
for saturation in saturation: 
    priors = saturation.sample_prior()
    saturation.plot_curve(parameters=priors, ax=ax)

[wd60622]

For instance, the tanhbasedline defaults priors seem to not support az.hdi with its defaults arguments. This can be seen by not have a confidence band in this plot: https://www.pymc-marketing.io/en/latest/api/generated/pymc_marketing.mmm.components.saturation.TanhSaturationBaselined.html