Question Answering on XQuAD (Language Subset)
83.9English QA ScoreBLOOM
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| BLOOMEvaluation Protocol=Understanding capability (Instruction identical to inputs)2024.06 | 83.9 | 83 | 79.9 | 27.4 | 79.2 | 22.8 | 82.7 | |
| BLOOMEvaluation Protocol=Generation capability (Instructions in English)2024.06 | 83.9 | 81.8 | 79.2 | 27.6 | 77.2 | 49.2 | 80.8 | |
| LLaMAEvaluation Protocol=Understanding capability (Instruction identical to inputs)2024.06 | 76.6 | 27.2 | 36.6 | 27.8 | 11.8 | 22.3 | 14.3 | |
| LLaMAEvaluation Protocol=Generation capability (Instructions in English)2024.06 | 76.6 | 66.3 | 42.9 | 38.1 | 24.2 | 40.7 | 30.8 | |
| ChatGPTEvaluation Protocol=Understanding capability (Instruction identical to inputs)2024.06 | 56 | 20.5 | 26.8 | 18.3 | 24.1 | 17.7 | 0.6 | |
| ChatGPTEvaluation Protocol=Generation capability (Instructions in English)2024.06 | 56 | 37.1 | 36.1 | 34.5 | 32 | 29.7 | 17.5 |