I’d like to predict a binomial output as a smooth interaction between two continuous variables that vary by a group term. Something like:

brm(y ~ t2(x1, x2) + (1 | group), data, family = binomial)

What is the proper way to specify this model in brms?