We've discussed this already for creating the "side" branches
But need to solidify how we will induce the LM to do this (strict definitions); and also whether we need to have some other special properties / structure (flexible causality?) so that we can do the downstream work of synthetic patient trajectories
We've discussed this already for creating the "side" branches
But need to solidify how we will induce the LM to do this (strict definitions); and also whether we need to have some other special properties / structure (flexible causality?) so that we can do the downstream work of synthetic patient trajectories