Toxicity Detection on Adv
59.74AccuracyInfICL
Evaluation results
| Method | Links | |
|---|---|---|
| InfICLBackbone=LLaMA3.1-8B2026.04 | 59.74 | |
| RewriteBackbone=LLaMA3.1-8B2026.04 | 59.67 | |
| RewriteBackbone=Qwen3-8B2026.04 | 59.41 | |
| DOPA*Backbone=LLaMA3.1-8B2026.04 | 59.28 | |
| DICLBackbone=Qwen3-8B2026.04 | 59.21 | |
| DOPA*Backbone=Qwen3-8B2026.04 | 59.08 | |
| DOPA*Backbone=Gemma2-2B2026.04 | 58.09 | |
| KNNBackbone=Qwen3-8B2026.04 | 58.03 | |
| InfICLBackbone=Qwen3-8B2026.04 | 57.96 | |
| InfICLBackbone=Gemma2-2B2026.04 | 57.5 | |
| RewriteBackbone=Qwen3-1.7B2026.04 | 57.5 | |
| RewriteBackbone=LLaMA3.2-3B2026.04 | 57.04 | |
| DrICLBackbone=Qwen3-8B2026.04 | 57.04 | |
| InfICLBackbone=Qwen3-1.7B2026.04 | 56.83 | |
| RandomBackbone=Qwen3-8B2026.04 | 56.71 | |
| DrICLBackbone=LLaMA3.1-8B2026.04 | 56.64 | |
| RandomBackbone=Gemma2-2B2026.04 | 56.51 | |
| DOPA*Backbone=LLaMA3.2-3B2026.04 | 56.45 | |
| DrICLBackbone=Gemma2-2B2026.04 | 56.45 | |
| DOPA*Backbone=Qwen3-1.7B2026.04 | 56.45 | |
| DrICLBackbone=LLaMA3.2-3B2026.04 | 56.18 | |
| DICLBackbone=LLaMA3.1-8B2026.04 | 55.99 | |
| DrICLBackbone=Qwen3-1.7B2026.04 | 55.86 | |
| RandomBackbone=LLaMA3.1-8B2026.04 | 55.66 | |
| RandomBackbone=LLaMA3.2-3B2026.04 | 55.2 | |
| InfICLBackbone=LLaMA3.2-3B2026.04 | 55.2 | |
| KNNBackbone=LLaMA3.1-8B2026.04 | 54.8 | |
| DICLBackbone=Qwen3-1.7B2026.04 | 54.54 | |
| DICLBackbone=LLaMA3.2-3B2026.04 | 53.88 | |
| DICLBackbone=Gemma2-2B2026.04 | 53.68 | |
| DOPA*Backbone=GPT2-xl2026.04 | 53.29 | |
| KNNBackbone=LLaMA3.2-3B2026.04 | 53.29 | |
| RandomBackbone=Qwen3-1.7B2026.04 | 52.89 | |
| RewriteBackbone=Gemma2-2B2026.04 | 51.84 | |
| DrICLBackbone=GPT2-xl2026.04 | 51.38 | |
| InfICLBackbone=GPT2-xl2026.04 | 50.33 | |
| RandomBackbone=GPT2-xl2026.04 | 50.2 | |
| KNNBackbone=Qwen3-1.7B2026.04 | 50.13 | |
| KNNBackbone=Gemma2-2B2026.04 | 47.89 | |
| KNNBackbone=GPT2-xl2026.04 | 47.5 | |
| DICLBackbone=GPT2-xl2026.04 | 47.37 | |
| RewriteBackbone=GPT2-xl2026.04 | 45.72 |