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

Single-hop Question Answering on Natural Questions (NQ) (test)

47.5EM

SE-Search-3B

-1.910.92523.7536.575Jan 25, 2026Jan 27, 2026Jan 29, 2026Jan 31, 2026Feb 2, 2026Feb 4, 2026Feb 6, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.02
47.5-
2026.02
46.7-
2026.02
44.4-
2026.02
43.6-
2026.02
43-
2026.02
42.7-
2026.02
42.1-
2026.02
42.1-
2026.02
39.7-
2026.02
36.5-
2026.01
35.9450.71
2026.02
34.8-
2026.01
34.3847.34
2026.01
33.5949.87
2026.01
32.0345.88
2026.01
32.0344.93
2026.01
30.4744.75
2026.02
24.9-
2026.01
24.2237.96
2026.01
24.2236.53
2026.02
23.8-
2026.02
22.6-
2026.02
21-
2026.01
16.4128.45
2026.01
11.7221.51
2026.02
11.1-
2026.02
10.6-
2026.01
5.1219.02
2026.01
3.1211.23
2026.01
2.348.9
2026.01
1.5613
2026.01
1.5615.97
2026.01
010.69