Find or carefully generate true parameters for mixture and bufferguts models and use for currently skipped comparisons in unit testing

If parameters are generated, we assume the models are correct. If we do so, it would be best practice to ask another group to independently validate the model estimates. Or at least validate the parameters with different implementations (we could use the original implementation and compare them to the pymob implementation)

Related: !6 (comment 14362)

I just saw again, that in the original GUTS Mixture paaper there is raw data provided in the SI together with their calibrated parameters. We could use this as the base for true parameters and testing: https://pubs.acs.org/doi/10.1021/acs.est.0c05282