Logical Reasoning Oversight on GGC
100AUCBlack
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| BlackModel=Gemma2026.06 | 100 | 0 | 0 | |
| BlackModel=GPT-4o-mini2026.06 | 100 | 0 | 0 | |
| BlackModel=Qwen2026.06 | 100 | 0 | 0 | |
| Pipeline 1Model=Gemma, Strategy=B3, Variant=pv12026.06 | 99.9 | 0 | 0 | |
| Pipeline 1Model=Gemma, Strategy=B4, Variant=pv32026.06 | 99.9 | 0 | 0 | |
| Pipeline 1Model=Qwen, Strategy=B4, Variant=pv32026.06 | 99.7 | 0 | 0.016 | |
| Pipeline 1Model=Qwen, Strategy=B2, Variant=pv32026.06 | 99.7 | 0 | 0.003 | |
| Pipeline 1Model=GPT-4o-mini, Strategy=B4, Variant=pv12026.06 | 99.3 | 0 | 0.072 | |
| Pipeline 1Model=GPT-4o-mini, Strategy=B4, Variant=pv22026.06 | 99 | 0 | 0.096 | |
| Pipeline 2Model=Gemma, Strategy=B3, Variant=pv42026.06 | 95.6 | 0.06 | 0.4 | |
| Pipeline 2Model=Gemma, Strategy=B3, Variant=pv22026.06 | 95.4 | 0.06 | 0.417 | |
| Pipeline 2Model=GPT-4o-mini, Strategy=B1, Variant=pv32026.06 | 93.3 | 0.213 | 0.481 | |
| Pipeline 2Model=GPT-4o-mini, Strategy=B3, Variant=pv12026.06 | 93.3 | 0.217 | 0.484 | |
| Pipeline 2Model=Qwen, Strategy=B2, Variant=pv42026.06 | 92.2 | 0.319 | 0.659 | |
| Pipeline 2Model=Qwen, Strategy=B4, Variant=pv12026.06 | 92.1 | 0.338 | 0.669 | |
| GreenModel=Gemma2026.06 | 91.6 | 0.408 | 0.704 | |
| GreenModel=Qwen2026.06 | 87.9 | 0.587 | 0.793 | |
| GreenModel=GPT-4o-mini2026.06 | 87.5 | 0.602 | 0.801 |