Checking reloo object for divergences in refitted models

I need to fit a large number of models and compare their elpd_loos, and I’d like to be able to have a script do all of this for me. When fitting the models themselves, it is easy enough to include a few lines of code to check for divergences and refit the model with a higher adapt_delta if there are any. You just use get_sampler_params().

Unfortunately, sometimes the original model converges without incident, yet divergences occur in one or more of the reloo refits. That’s when it becomes necessary to increase adapt_delta for the reloo refits rather than for the original fit. I imagine this can be done by adding reloo_args = list(adapt_delta = HigherValue) to the loo() call.

But when I’m doing all of this with a script that loops through a large set of candidate models, how do I check a reloo object for divergences? It appears get_sampler_params() is unable to retrieve information from refits done within a loo(x, reloo = TRUE) call. Can it be done?