metadata
dataset_info:
features:
- name: sentence
dtype: string
- name: template
dtype: string
- name: person
dtype: string
- name: gender
dtype: string
- name: race
dtype: string
- name: emotion
dtype: string
- name: emotion word
dtype: string
splits:
- name: train
num_bytes: 318577
num_examples: 2000
- name: same_iid
num_bytes: 321274
num_examples: 2000
- name: higher_correlation
num_bytes: 323884
num_examples: 2000
- name: reduced_correlation
num_bytes: 321595
num_examples: 2000
- name: flipped
num_bytes: 318781
num_examples: 2000
download_size: 154291
dataset_size: 1604111
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: same_iid
path: data/same_iid-*
- split: higher_correlation
path: data/higher_correlation-*
- split: reduced_correlation
path: data/reduced_correlation-*
- split: flipped
path: data/flipped-*