Hierarchy Transformer (HiT) is a framework that enables transformer encoder-based language models (LMs) to learn hierarchical structures in hyperbolic space.
