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

Open Domain Question Answering on Natural Questions (NQ)

60.7Exact Match (EM)

DeepSeek-v3

1.773617.071832.3747.6682Jul 2, 2020Jun 14, 2021May 28, 2022May 11, 2023Apr 22, 2024Apr 5, 2025Mar 19, 2026
Updated 26d ago

Evaluation Results

MethodLinks
2026.03
60.7
2026.03
60.3
2026.03
59.4
2026.03
59
2026.03
57.6
2026.03
57.6
2026.03
56.2
2026.03
56
2026.03
55.1
2026.03
55
2026.03
53.8
2026.03
53.6
2020.07
51.4
2020.09
51.4
2026.03
50.3
2026.03
50.3
2026.03
48.3
2026.03
48.3
2020.07
48.2
2026.03
46.6
2020.09
45.3
2020.07
44.5
2020.09
44.5
2025.12
44.1
2020.09
43.8
2020.07
42.5
2020.09
42.2
2020.09
41.8
2020.07
41.5
2020.09
41.5
2025.12
41.4
2020.07
40.4
2020.09
40.4
2026.03
39.5
2020.09
38.1
2025.12
38.1
2025.12
37.8
2020.09
37.7
2020.07
36.6
2020.09
36.6
2025.12
36.4
2025.12
35.6
2020.09
35.3
2025.12
34.9
2025.12
34.8
34.7
2020.09
34.5
2020.09
33.3
2025.12
32.7
2020.09
32.6
31.7
2020.07
31.3
2020.07
29.9
2020.09
29.9
2025.12
29.3
2020.07
28.8
2025.12
28.8
2020.09
28.1
2025.12
24.9
2025.12
22.4
2025.12
15.1
2025.12
13.4
2026.03
11.83
2026.03
9.48
2025.12
9.4
2026.03
8.6
2026.03
7.56
2026.03
6.9
2026.03
5.9
2026.03
5.48
2026.03
4.96
2025.12
4.8
2026.03
4.32
2026.03
4.04