Hate Speech Detection
Hate speech detection is the task of detecting if communication such as text, audio, and so on contains hatred and or encourages violence towards a person or a group of people. This is usually based on prejudice against 'protected characteristics' such as their ethnicity, gender, sexual orientation, religion, age et al. Some example benchmarks are ETHOS and HateXplain. Models can be evaluated with metrics like the F-score or F-measure.
Papers
Showing 1–10 of 507 papers
All datasetsEthos BinaryHateXplainEthos MultiLabelWaseem et al., 2018AbusEvalAutomatic Misogynistic IdentificationHateMMHatEvalOffensEval 2019ToLD-Brbajer_danish_misogynyDKhate
Benchmark Results
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | BiLSTM + static BE | F1-score | 0.8 | — | Unverified |
| 2 | BERT | F1-score | 0.79 | — | Unverified |
| 3 | BiLSTM+Attention+FT | F1-score | 0.77 | — | Unverified |
| 4 | OPT-175B (few-shot) | F1-score | 0.76 | — | Unverified |
| 5 | CNN+Attention+FT+GV | F1-score | 0.74 | — | Unverified |
| 6 | OPT-175B (one-shot) | F1-score | 0.71 | — | Unverified |
| 7 | OPT-175B (zero-shot) | F1-score | 0.67 | — | Unverified |
| 8 | SVM | F1-score | 0.66 | — | Unverified |
| 9 | Random Forests | F1-score | 0.64 | — | Unverified |
| 10 | Davinci (zero-shot) | F1-score | 0.63 | — | Unverified |