EPFL's New Large Language Model for Medical Knowledge

A <i>13B parameter model</i> and you claim careful curation by experts, do not name the training data sources, only rely on open data, and then, like everyone else, want researchers to work (for free) to improve your model?