metadata
dataset_info:
features:
- name: id
dtype: string
- name: prompt
dtype: string
- name: gt_is_correct
dtype: bool
- name: gt_token_count
dtype: int64
- name: source_model
dtype: string
- name: retrieved_anchors
sequence: string
splits:
- name: train
num_bytes: 18614899
num_examples: 3276
download_size: 6634852
dataset_size: 18614899
configs:
- config_name: default
data_files:
- split: train
path: data/train-*