Open-ended QA on Open-ended QA
99.9PrecisionPost-trained
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Post-trainedBackbone=Qwen2026.02 | 99.9 | 0.659 | 8.6 | |
| Post-trainedBackbone=Gemma2026.02 | 99.4 | 0.771 | 9.6 | |
| Proxy-SoupBackbone=Gemma2026.02 | 99.3 | 0.967 | 12.1 | |
| Post-trainedBackbone=Llama2026.02 | 98.6 | 1.442 | 20.4 | |
| Proxy-SoupBackbone=Llama2026.02 | 98.1 | 1.76 | 28.5 | |
| SLRBackbone=Qwen2026.02 | 97.8 | 1.583 | 22.6 | |
| SLRBackbone=Gemma2026.02 | 97.8 | 1.827 | 30.7 | |
| Proxy-SoupBackbone=Qwen2026.02 | 97.7 | 0.919 | 12.4 | |
| SLRBackbone=Llama2026.02 | 96.8 | 2.322 | 46.1 |