| language: | |
| - en | |
| 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 | |