Offensive Statement Classification on COBRACORPUS-CF
71.4AccuracyCHARM
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| CHARMmodel_size=XXL, context_availability=With Context2023.06 | 71.4 | 96.5 | 64.2 | 77.1 | |
| CHARMmodel_size=XL, context_availability=With Context2023.06 | 66.5 | 98.84 | 60 | 74.7 | |
| GPT-3.5model_version=text-davinci-003 Jan 13th 20222023.06 | 55.2 | 99.4 | 52.7 | 68.9 | |
| All Toxicstrategy=Predicting every statement as toxic2023.06 | 50 | 100 | 50 | 67.8 | |
| CHARMmodel_size=XL, context_availability=Without Context (WoC)2023.06 | 50 | 72.3 | 50 | 59.1 |