Ethics Question Answering on EthicsQA
83.2AccuracyDeepSeek-R1
Evaluation results
| Method | Links | |
|---|---|---|
| DeepSeek-R1Model Size=671B, Mode=Thinking2026.02 | 83.2 | |
| MirrorModel Size=8B, Mode=No-Thinking2026.02 | 76.07 | |
| Qwen3Model Size=32B, Mode=No-Thinking2026.02 | 73.8 | |
| Qwen3Model Size=32B, Mode=Thinking2026.02 | 73.8 | |
| GPT-4.1Mode=No-Thinking2026.02 | 73.2 | |
| MirrorModel Size=8B, Mode=Thinking2026.02 | 69.33 | |
| Qwen3Model Size=14B, Mode=No-Thinking2026.02 | 67.4 | |
| Qwen3Model Size=14B, Mode=Thinking2026.02 | 67.27 | |
| Qwen3Model Size=8B, Mode=No-Thinking2026.02 | 65.4 | |
| Qwen3Model Size=8B, Mode=Thinking2026.02 | 63.6 |