Legal Reasoning on CaseHold (Accuracy)
83.13Accuracy (CaseHold)XPERT-DeepSeek
Evaluation Results
| Method | Links | |
|---|---|---|
| XPERT-DeepSeek#Params=270M2026.05 | 83.13 | |
| XPERT-OLMoE#Params=270M2026.05 | 83.11 | |
| XPERT-OLMoE#Params=570M2026.05 | 82.96 | |
| XPERT-OLMoE#Params=480M2026.05 | 82.9 | |
| Distillation#Params=270M2026.05 | 82.54 | |
| XPERT-DeepSeek#Params=570M2026.05 | 82.54 | |
| XPERT-DeepSeek#Params=480M2026.05 | 82.12 | |
| XPERT-OLMoE#Params=391M2026.05 | 81.5 | |
| Distillation#Params=480M2026.05 | 81.35 | |
| Distillation#Params=570M2026.05 | 81.35 | |
| XPERT-DeepSeek#Params=391M2026.05 | 81.1 | |
| Scratch#Params=480M2026.05 | 80.97 | |
| Scratch#Params=270M2026.05 | 80.93 | |
| Distillation#Params=391M2026.05 | 80.48 | |
| Scratch#Params=570M2026.05 | 80.16 | |
| Scratch#Params=391M2026.05 | 78.5 |