| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Question Answering | NQ (Natural Questions) in-domain (test) | Exact Match48.8 | 29 | |
| Single-hop Question Answering | NQ (Natural Questions) (test) | Accuracy45.3 | 21 | |
| Single-Hop Question Answering | NQ (Natural Questions) 2018 Wikipedia dump (dev) | Accuracy50.2 | 14 | |
| Question Answering | NQ (Natural Questions) December 2018 Wikipedia dump (test) | EM36.9 | 14 | |
| Question Retrieval | NQ (Natural Questions) (full) | Retrieval Accuracy40.9 | 12 | |
| Question Answering | NQ (Natural Questions) (out-of-domain) | F1 Score14.3 | 3 |