Red-teaming on Religious Discrimination principle v1 (test)
5.32Mean Best Category ScoreQCI
Evaluation Results
| Method | Links | |
|---|---|---|
| QCITarget Model=Llama-3.1-8B-Instruct2026.02 | 5.32 | |
| CRLTarget Model=Llama-3.1-8B-Instruct2026.02 | 4.84 | |
| QCITarget Model=Qwen3-30B-A3B-Instruct-25072026.02 | 4.14 | |
| QCITarget Model=GPT-4.1-Mini2026.02 | 2.82 | |
| QCITarget Model=Gemma3-12B-IT2026.02 | 2.66 | |
| CRLTarget Model=GPT-4.1-Mini2026.02 | 2.57 | |
| CRLTarget Model=Gemma3-12B-IT2026.02 | 2.26 | |
| CRLTarget Model=Qwen3-30B-A3B-Instruct-25072026.02 | 2.22 | |
| RSTarget Model=Llama-3.1-8B-Instruct2026.02 | -0.9 | |
| RSTarget Model=Qwen3-30B-A3B-Instruct-25072026.02 | -1.29 | |
| RSTarget Model=Gemma3-12B-IT2026.02 | -1.62 | |
| RSTarget Model=GPT-4.1-Mini2026.02 | -1.71 |