Question Answering on AGIEval (Vanilla Accuracy, AbsAcc)
43.92Vanilla AccuracyMeanLearn
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MeanLearnSize=13B2024.03 | 43.92 | 30.01 | |
| Orca-2Size=13B2024.03 | 43.75 | 50 | |
| Orca-2Size=7B2024.03 | 38.57 | 26.27 | |
| MeanLearnSize=8B2024.03 | 38.53 | 26.57 | |
| MeanLearnSize=7B2024.03 | 38.15 | 25.27 | |
| LLaMA-3Size=8B2024.03 | 36.33 | 23.78 | |
| LLaMA-2Size=13B2024.03 | 34.41 | 21.01 | |
| VicunaSize=13B2024.03 | 33.23 | 19.7 | |
| WizardLMSize=13B2024.03 | 33.23 | 16.96 | |
| VicunaSize=7B2024.03 | 32.56 | 19.9 | |
| LLaMA-2Size=7B2024.03 | 27.89 | 16.09 | |
| GPT-3.5Size=>20B2024.03 | 25.84 | 14.48 | |
| WizardLMSize=7B2024.03 | 22.27 | 12.94 | |
| PaLM-2Size=>20B2024.03 | 15.83 | 9.59 |