Adam Smith [tweets](https://twitter.com/anhsmith/status/1050560008104996864): "How about [LKJ](https://www.psychstatistics.com/2014/12/27/d-lkj-priors/)?" Fortunately there's already a density function for it in the [rethinking](https://github.com/rmcelreath/rethinking) package: https://rdrr.io/github/rmcelreath/rethinking/man/dlkjcorr.html