metadata
dataset_info:
features:
- name: question
dtype: string
- name: prob
dtype: float64
- name: correct_count
dtype: int64
- name: total_samples
dtype: int64
- name: question_hash
dtype: string
- name: answer_idx
dtype: string
splits:
- name: train
num_bytes: 8914365
num_examples: 11448
download_size: 4805216
dataset_size: 8914365
configs:
- config_name: default
data_files:
- split: train
path: data/train-*