Knowledge MCQ on Belebele
71.43Accuracyend-to-end
Evaluation Results
| Method | Links | |
|---|---|---|
| end-to-endModel=GPT-4o-mini2026.06 | 71.43 | |
| context-aware cascadeModel=GPT-4o-mini2026.06 | 58.24 | |
| context-aware cascadeModel=Llama-3.1-8B-Instruct2026.06 | 58.18 | |
| standard cascadeModel=Llama-3.1-8B-Instruct2026.06 | 56.89 | |
| end-to-endModel=Llama-3.1-8B-Instruct2026.06 | 53.5 | |
| standard cascadeModel=GPT-4o-mini2026.06 | 53.5 | |
| end-to-endModel=Mistral-7B-Instruct-v0.32026.06 | 33.69 | |
| context-aware cascadeModel=Mistral-7B-Instruct-v0.32026.06 | 25.92 | |
| standard cascadeModel=Mistral-7B-Instruct-v0.32026.06 | 24.6 |