A massive open repository of web crawl data used to train most large language models.