---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: question
    dtype: string
  - name: answer
    dtype: string
  splits:
  - name: train
    num_bytes: 1473584741.0
    num_examples: 200
  download_size: 1276133511
  dataset_size: 1473584741.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---
**Please see paper & code for more information:**
- https://github.com/mbzuai-oryx/KITAB-Bench
- https://arxiv.org/abs/2502.14949