Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

General Question Answering on Natural Questions (NQ) (test val)

55.8EM

DeepControl

0.1614.60529.0543.495Feb 2, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.02
55.847.9
2026.02
53.341.1
2026.02
4843.1
2026.02
40.630.3
2026.02
39.338.5
2026.02
3634.8
2026.02
34.930.4
2026.02
34.827
2026.02
34.132.5
2026.02
31.820.7
2026.02
29.727.6
2026.02
29.426.5
2026.02
2727.1
2026.02
24.917.6
2026.02
23.825.5
2026.02
22.622.9
2026.02
22.423.9
2026.02
2122.4
2026.02
15.120.6
2026.02
13.418.1
2026.02
11.118.1
2026.02
10.613.4
2026.02
4.810.6
2026.02
2.31.5