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

Single-Hop Question Answering on NQ (EM)

51.7Exact Match (EM)

Tree-GRPO

5.31617.35829.441.442Sep 25, 2025Nov 3, 2025Dec 12, 2025Jan 20, 2026Feb 28, 2026Apr 8, 2026May 18, 2026
Updated 14d ago

Evaluation Results

MethodLinks
2025.09
51.7
2025.09
51.3
2025.09
50.7
2026.05
50.2
2026.05
50
2026.05
49.5
2026.03
48.4
2026.05
48.4
2025.09
48.1
2026.05
48
2025.09
47.7
2025.09
47
2026.05
46.9
2025.09
46.8
2026.03
46.7
2026.05
45.9
2025.09
45.8
2025.09
45.5
2026.05
45
2025.09
44.4
2025.09
43.6
2026.05
43.6
2026.05
43.5
2025.09
43
2025.09
41.2
2025.09
39.4
2026.03
39.3
2026.05
39.3
2026.05
39.3
2026.03
38.5
2026.03
38.5
2025.09
36.8
2025.09
36.1
2026.03
34.9
2026.03
34.8
2025.09
34.7
2026.03
34.1
2025.09
30.6
2026.05
27.4
2025.09
24.2
2025.09
23.9
2026.03
23.8
2025.09
23.8
2026.03
22.4
2025.09
21.1
2025.09
19.8
2026.05
19.4
2025.09
16.2
2026.03
15.1
2025.09
15.1
2026.03
13.4
2025.09
13.4
2026.05
12.8
2026.05
11.6
2026.03
11.1
2026.03
10.6
2025.09
10.6
2025.09
10.2
2025.09
9.5
2025.09
7.1