|
|
--- |
|
|
dataset_info: |
|
|
features: |
|
|
- name: text |
|
|
dtype: string |
|
|
- name: label |
|
|
dtype: string |
|
|
splits: |
|
|
- name: train |
|
|
num_bytes: 1583633 |
|
|
num_examples: 11240 |
|
|
- name: valid |
|
|
num_bytes: 144044 |
|
|
num_examples: 1000 |
|
|
- name: test |
|
|
num_bytes: 141548 |
|
|
num_examples: 1000 |
|
|
download_size: 1156986 |
|
|
dataset_size: 1869225 |
|
|
configs: |
|
|
- config_name: default |
|
|
data_files: |
|
|
- split: train |
|
|
path: data/train-* |
|
|
- split: valid |
|
|
path: data/valid-* |
|
|
- split: test |
|
|
path: data/test-* |
|
|
--- |
|
|
|
|
|
# Abuseval |
|
|
|
|
|
Note: this is not the official valid/test split. |
|
|
|
|
|
Abusive language detection, with three labels: NOTABU (not abusive), EXP (explicit), IMP (implicit) |
|
|
|
|
|
English tweets. User names, urls, etc have been replaced. |
|
|
See: https://github.com/tommasoc80/AbuseEval, https://github.com/joeykay9/offenseval |
|
|
|
