Question Answering on WebQS (test)
27.7AccuracyLONGLLAMA
Evaluation Results
| Method | Links | |
|---|---|---|
| LONGLLAMAModel size=7B, Context length=8K2023.07 | 27.7 | |
| LONGLLAMAModel size=7B, Context length=6K2023.07 | 27.2 | |
| LONGLLAMAModel size=7B, Context length=4K2023.07 | 26.4 | |
| LONGLLAMAModel size=7B, Context length=2K2023.07 | 25.5 | |
| LONGLLAMAModel size=3B, Context length=8K2023.07 | 22.4 | |
| LONGLLAMAModel size=3B, Context length=6K2023.07 | 22.2 | |
| LONGLLAMAModel size=3B, Context length=4K2023.07 | 21.4 | |
| LONGLLAMAModel size=3B, Context length=2K2023.07 | 21.2 |