Recovering latent nonlinear responses

Hi all,

Apologies if I’ve missed this in the docs.

Short of manually calculating them, is there a built-in method to recover/predict the parameters b0 & b1 for a more complicated version of a model fitted using the following formula?

# Simplified for this question
bf_fit<-bf(y ~  b0 + b1*(x1 + b1),
                    b0 ~ 1 + (1 | p),
                    b1 ~ x2 + (1 | p),

What Have I’ve tried?

predict.brmsfit with e.g. resp=‘b0’ seems to predict y.


Try fitted with the nlpar="b0" argument.

Perfect, thanks! I’ve been running myself round in circles.