C4
Dataset
Web Corpus
Google's cleaned Common Crawl — the reproducible baseline.
You're looking at a dataset record: a training-data collection — what's in it and which models learned from it.
The Colossal Clean Crawled Corpus (2019, for T5) applied documented filters to Common Crawl and released the recipe, becoming the standard public pre-training baseline and the corpus of record for studying what models actually read.
How this connects to the rest of the stack
This record's links at a glance — click a group to jump to it.
Related
Sources
Record reviewed as of 2026-01-15 — see Methodology.