metadata
dataset_info:
features:
- name: id
dtype: int64
- name: question_id
dtype: int64
- name: comment
dtype: string
- name: stance_label
dtype: string
- name: author
dtype: string
- name: label
dtype: int64
- name: topic
dtype: string
- name: language
dtype: string
- name: question_en
dtype: string
- name: question_de
dtype: string
- name: question_fr
dtype: string
- name: question_it
dtype: string
splits:
- name: train
num_bytes: 37265285
num_examples: 45640
- name: validation
num_bytes: 3144692
num_examples: 3926
- name: test
num_bytes: 13762384
num_examples: 17705
download_size: 8492026
dataset_size: 54172361
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*
Data cloned from https://huggingface.co/datasets/ZurichNLP/x_stance, where you will find more information.
One exception is that the label column is now a numerical value with a fixed mapping to the stance_label column, and the numerical_label column is dropped.