Hi @franzsf,
I am dealing with the same issue. In trying to reduce divergent transitions for a model discussed here, it was suggested that modeling correlations among population-level effects and group-level effects might help. However, I have not been able to implement the multi_normal() prior using brms.
Have you found a way to specify the multi-normal prior in the NL formula syntax? Would appreciate your feedback.
Thanks,
Meghna