@article{hierarchicalautoregressivetransformers, title = {Hierarchical Autoregressive Transformers: Combining Byte- and Word-Level Processing for Robust, Adaptable Language Models}, author = {Pit Neitemeier and Björn Deiseroth and Constantin Eichenberg and Lukas Balles}, year = {2025}, eprint = {2501.10322}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2501.10322v2}, }