metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: category
dtype: string
- name: extra_info
struct:
- name: dataset
dtype: string
- name: ground_truth
struct:
- name: answer_letter
dtype: string
- name: answer_text
dtype: string
- name: high_level_domain
dtype: string
- name: record_id
dtype: string
- name: subdomain
dtype: string
- name: options
list: string
- name: scoring_function
dtype: string
- name: variant
dtype: string
splits:
- name: train
num_bytes: 724290
num_examples: 448
download_size: 268464
dataset_size: 724290
configs:
- config_name: default
data_files:
- split: train
path: data/train-*