metadata
dataset_info:
features:
- name: dimension_name
dtype: string
- name: dimension_values
dtype: string
- name: dimension_description
dtype: string
- name: question
dtype: string
- name: why_differ
dtype: string
- name: how_subtle
dtype: string
- name: sampled_value
dtype: string
- name: num_distractors
dtype: int64
- name: ground_truth_persona
dtype: string
- name: distractor_personas
dtype: string
splits:
- name: train
num_bytes: 21185
num_examples: 10
download_size: 22431
dataset_size: 21185
configs:
- config_name: default
data_files:
- split: train
path: data/train-*