Hate speech classification on HateXplain (test)
72Macro-F1Llama_Full
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Llama_FullTraining Paradigm=Full, Model Backbone=Llama, Data%=100%2025.09 | 72 | 100 | |
| T5_FullTraining Paradigm=Full, Model Backbone=T5, Data%=100%2025.09 | 72 | 100 | |
| ModernBERTTraining Paradigm=Full, Model Backbone=ModernBERT, Data%=100%2025.09 | 70 | 97 | |
| SMARTER (Llama_DPO-256)Training Paradigm=DPO, K shots=256, Model Backbone=Llama, Data%=6%2025.09 | 64 | 89 | |
| SMARTER (T5_DPO-256)Training Paradigm=DPO, K shots=256, Model Backbone=T5, Data%=6%2025.09 | 62 | 86 | |
| GPT-5-chatTraining Paradigm=16-shot ICL, K shots=16, Model Backbone=GPT-5-chat2025.09 | 62 | 86 | |
| GPT-5-chatTraining Paradigm=Zero-shot, K shots=0, Model Backbone=GPT-5-chat2025.09 | 56 | 78 | |
| GPT-4.1Training Paradigm=Zero-shot, K shots=0, Model Backbone=GPT-4.12025.09 | 55 | 76 | |
| Qwen-32BTraining Paradigm=16-shot ICL, K shots=16, Model Backbone=Qwen-32B2025.09 | 55 | 76 | |
| Qwen-32BTraining Paradigm=Zero-shot, K shots=0, Model Backbone=Qwen-32B2025.09 | 54 | 75 | |
| GPT-4.1Training Paradigm=16-shot ICL, K shots=16, Model Backbone=GPT-4.12025.09 | 52 | 72 | |
| GPT-4o-miniTraining Paradigm=Zero-shot, K shots=0, Model Backbone=GPT-4o-mini2025.09 | 50 | 69 | |
| GPT-4o-miniTraining Paradigm=16-shot ICL, K shots=16, Model Backbone=GPT-4o-mini2025.09 | 29 | 40 |