metadata

language:
  - ar
task_categories:
  - multiple-choice
size_categories:
  - n<1K
viewer: true
license: apache-2.0

EXAMs

You can find details of the dataset in this post:https://arxiv.org/pdf/2308.16149.pdf

About this Arabic dataset

We only took the Arabic part of the dataset,which contains 562 data.
We then extracted five from each category based on the task domain as a few shot data.