Ethics Question Answering on EthicsQA ER
0.8212AccuracyDeepSeek-R1
Evaluation results
| Method | Links | |
|---|---|---|
| DeepSeek-R1Model Size=671B, Mode=Thinking2026.02 | 0.8212 | |
| Qwen3Model Size=32B, Mode=Thinking2026.02 | 0.6769 | |
| MirrorModel Size=8B, Mode=No-Thinking2026.02 | 0.6606 | |
| Qwen3Model Size=32B, Mode=No-Thinking2026.02 | 0.6473 | |
| MirrorModel Size=8B, Mode=Thinking2026.02 | 0.6294 | |
| GPT-4.1Mode=No-Thinking2026.02 | 0.6221 | |
| Qwen3Model Size=14B, Mode=Thinking2026.02 | 0.5968 | |
| Qwen3Model Size=14B, Mode=No-Thinking2026.02 | 0.5786 | |
| Qwen3Model Size=8B, Mode=Thinking2026.02 | 0.5667 | |
| Qwen3Model Size=8B, Mode=No-Thinking2026.02 | 0.5395 |