172 DocLLM: A layout-aware generative language model for multimodal document understanding · 9 authors 24