Aya Dataset
Human instructions, Wolof included
Official site consultedAI: datasets
Last verified: 1 October 2026
The Aya Dataset is a human-annotated instruction dataset: over 204,000 prompt-completion pairs in 65 languages, including Wolof (code wol), at about 140 MB. It is under Apache 2.0, usable for any purpose.
Services
- Original annotations and re-annotations
- Annotator demographics data
Access requirements
Apache 2.0 license, academic or commercial use.
Sources
- Aya Dataset (new tab), verified on 1 October 2026Official site consulted