metadata
dataset_info:
features:
- name: question
dtype: string
- name: optionA
dtype: string
- name: optionB
dtype: string
- name: optionC
dtype: string
- name: optionD
dtype: string
- name: solution
dtype: string
splits:
- name: train
num_bytes: 5813
num_examples: 50
download_size: 7834
dataset_size: 5813
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
license: apache-2.0
task_categories:
- question-answering
language:
- cs
size_categories:
- n<1K
Synthetic dataset of 50 Czech A/B/C/D questions for quick evaluation how the LMM understands Czech and Czech specific knowledge. Draft generated with the help of GPT4 and later manualy validated and edited.
from datasets import load_dataset