Our preprint Quantizing Recurrent Language Models Is Easier Than You Think is now public! See the paper and blog post.