metadata
dataset_info:
features:
- name: choices
sequence: string
- name: example
dtype: string
- name: gold
dtype: int64
- name: correct_answer_count
dtype: int64
- name: acc_norm
dtype: int64
- name: model
sequence: string
- name: acc_list
sequence: int64
- name: acc_norm_list
sequence: int64
- name: acc_total
sequence: float64
- name: predicted
sequence: int64
- name: predictions
sequence:
sequence: float64
splits:
- name: train
num_bytes: 6975552078
num_examples: 10019
download_size: 1781172843
dataset_size: 6975552078
configs:
- config_name: default
data_files:
- split: train
path: data/train-*