metadata
dataset_info:
config_name: multilabel
features:
- name: comment
dtype: string
- name: violence
dtype: float64
- name: directed_vs_generalized
dtype: float64
- name: gender
dtype: float64
- name: race
dtype: float64
- name: national_origin
dtype: float64
- name: disability
dtype: float64
- name: religion
dtype: float64
- name: sexual_orientation
dtype: float64
splits:
- name: train
num_bytes: 79108
num_examples: 433
download_size: 43700
dataset_size: 79108
configs:
- config_name: multilabel
data_files:
- split: train
path: multilabel/train-*
source_datasets:
- iamollas/ethos
ethos
ETHOS multi-label: hateful comments with the share of raters who saw each aspect (violence, target, grounds).
Only the source's single split is provided.
Original data: iamollas/ethos. Repackaged as parquet for tasksource by scripts/upload_repackaged.py.