Why teams choose it
Give your AI pipeline a broad, clean test. Corpus One has 999 useful knowledge records with clear sources, fixed splits, and proof for every file.
Use it to
- Test a training pipeline
- Build search and embeddings
- Check data from start to finish
Know before you use it
- Too small for broad pretraining
- Carries bias from the source catalog
- Includes some sensitive history and social topics
Lineage
Foundation of the C1–C4 line
Clear rights notes
The records use pinned Wikidata CC0 data. The catalog used to choose them has separate credit notes.