The only resource I found explaining the default priors in brms is its manual (newest version, updated 03/14/2021) for function set_prior().
For the intercept, the manual does not specify how the default prior is selected. When I fit a generalized linear mixed-effects model (GLMM), it appeared to be Normal(0, scale), but I am not sure how the scale was determined.
For standard deviations, such as the ones for group-level random effects and for the Gaussian auxiliary parameter (residual standard deviation), the default prior is a half student-t distribution t(3, 0, scale). The manual describes that the scale "depends on the standard deviation of the response after applying the link function. Minimally, the scale parameter is 2.5". However I cannot find any details on this dependency.
In comparison, the default priors implemented by another package rstanarm has clear descriptions on their default prior choices especially on how the scales are determined.
I wonder if anyone knows more about (1) the default prior distribution for the intercept, (2) how the scale is determined for the half student-t for the standard deviations in brms.
