Download or look at the dataset on GitHub.
Accessing the literacy dataset
We’re creating the benchmark data for our literacy evaluators by adding expert text-complexity annotations to part of the CLEAR (CommonLit Ease of Readability) Corpus.
This first release focuses on Grades 3 and 4, with educator-scored examples in areas like sentence structure and vocabulary. It’s designed as a practical resource for researchers, developers, and the learning science community interested in exploring readability and text complexity more deeply.
Check back for updates — we’re continually expanding our scoring to include more grades and new qualitative dimensions from Student Achievement Partners’ SCASS Rubric for Informational Text.
Learn more about the literacy dataset.