Bias Detection on Implicit Toxicity (test)
81AccuracyBiasGuard
Evaluation results
| Method | Links | ||
|---|---|---|---|
| BiasGuard2025.04 | 81 | 1.25 | |
| GPT-4oModel Category=Rule-based LLMs as Judges2025.04 | 75 | 5.6 | |
| Llama-3-8B-itModel Category=Rule-based LLMs as Judges2025.04 | 71.1 | 3 | |
| DeepSeek-R1-32BModel Category=Rule-based LLMs as Judges2025.04 | 62.44 | 8.06 | |
| GPT-4oModel Category=LLMs as Judges2025.04 | 54.25 | 5 | |
| Azure APIModel Category=Bias Classifier2025.04 | 49.11 | 8.27 | |
| DeepSeek-R1-32BModel Category=LLMs as Judges2025.04 | 45 | 21.6 | |
| ToxigenModel Category=Bias Classifier2025.04 | 41.3 | 4.35 | |
| Llama-Guard-3Model Category=Bias Classifier2025.04 | 34.6 | 0.25 | |
| Llama-3-8B-itModel Category=LLMs as Judges2025.04 | 30.85 | 3.45 | |
| ShieldGemmaModel Category=Bias Classifier2025.04 | 27.3 | 1 | |
| Moderation APIModel Category=Bias Classifier2025.04 | 25.5 | 0.1 |