Short Answer Question Answering on Natural Questions (NQ) (test)
67.4F1 ScoreAGAR
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| AGARBackbone=InternVL3.5-8B2026.06 | 67.4 | — | — | — | |
| RikiNet2020.09 | 61.3 | 67.6 | 56.1 | — | |
| Cluster-Former2020.09 | 60.9 | 62.1 | 59.8 | — | |
| VEA-ConBackbone=InternVL3.5-8B2026.06 | 59.9 | — | — | — | |
| UniformBackbone=InternVL3.5-8B2026.06 | 59.6 | — | — | — | |
| RandomBackbone=InternVL3.5-8B2026.06 | 59.3 | — | — | — | |
| Plain VQABackbone=InternVL3.5-8B2026.06 | 59.1 | — | — | — | |
| VEA-BrBackbone=InternVL3.5-8B2026.06 | 59 | — | — | — | |
| BigBird-ETC-large2020.09 | 57.9 | 63.7 | 53 | — | |
| ETC2020.04 | 57.86 | — | — | 18 | |
| AGLABackbone=InternVL3.5-8B2026.06 | 53 | — | — | — | |
| VARBackbone=InternVL3.5-8B2026.06 | 52.4 | — | — | — | |
| BERT jointpre-training=SQuAD 1.1 fine-tuned2019.01 | 52.1 | 63.8 | 44 | — | |
| DocumentQA2019.01 | 35.1 | 40.6 | 31 | — | |
| CGRBackbone=InternVL3.5-8B2026.06 | 33.6 | — | — | — | |
| DecAtt + DocReader2019.01 | 31.5 | 31.9 | 31.1 | — |