Amharic preference and evaluation data
Published
We do not yet have a Amharic preference set. Here is the exact format it would take, the English sample that proves the pipeline, and how to commission one with us.
The format, schema, provenance discipline, and privacy posture are identical to our human-preference dataset. No Amharic rows exist yet; this page describes what we can build with you.
Help build this
Want to help create it? Contributors can open a tab or tap to train in Amharic.