"Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer" by Google researchers introduces the T5 model, which converts NLP tasks into a unified text-to-text format. The paper also introduces the Colossal Clean Crawled Corpus (C4) for pre-training [arxiv.org/abs/1910.10683]. Access the full paper, published in JMLR, at arXiv.
