Reading Comprehension on XQUAD
90.3Accuracy (ZH)ORACLE
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| ORACLEBase Model=Llama3.3-70B2026.04 | 90.3 | 91.9 | 92.4 | 88.4 | 90.8 | |
| NATIVEBase Model=Llama3.3-70B2026.04 | 86.1 | 87.4 | 88.4 | 84.5 | 86.6 | |
| MLP (Ours)Base Model=Llama3.3-70B2026.04 | 86.1 | 87.3 | 88.8 | 84.5 | 86.7 | |
| XGBOOST (Ours)Base Model=Llama3.3-70B2026.04 | 85.8 | 87.8 | 89.1 | 85.2 | 87 | |
| TRANSLATEBase Model=Llama3.3-70B2026.04 | 83.9 | 86.6 | 88.3 | 80.1 | 84.7 |