Natural Language Inference on NLI
91.2AccuracyEA
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| EABase Model=GLM-4-Plus2025.12 | 91.2 | — | |
| EABase Model=Doubao-1.5-pro-32k2025.12 | 91.14 | — | |
| EABase Model=Qwen-Max2025.12 | 90.92 | — | |
| GLM-4-Plus2025.12 | 90.56 | — | |
| EABase Model=Hunyuan-Turbos2025.12 | 90.18 | — | |
| Qwen-Max2025.12 | 90.1 | — | |
| Doubao-1.5-pro-32k2025.12 | 89.98 | — | |
| Hunyuan-Turbos2025.12 | 88.62 | — | |
| ABMLL-MetaICLModel=LLAMA3, ICL setting=Yes2025.08 | 85.2 | 20.7 | |
| MetaICLModel=QWEN2, ICL setting=Yes2025.08 | 83.8 | 24.6 | |
| Reptile-MetaICLModel=LLAMA3, ICL setting=Yes2025.08 | 83.6 | 28.1 | |
| MetaICLModel=LLAMA3, ICL setting=Yes2025.08 | 83.3 | 24.6 | |
| Reptile-MetaICLModel=QWEN2, ICL setting=Yes2025.08 | 83 | 25.2 | |
| ABMLL-MetaICLModel=QWEN2, ICL setting=Yes2025.08 | 82.6 | 22.7 |