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

Multi-Hop Question Answering on HotpotQA (EM)

50.2Exact Match (EM)

Tree-GRPO

14.42423.7123342.288Sep 25, 2025Oct 27, 2025Nov 29, 2025Dec 31, 2025Feb 2, 2026Mar 6, 2026Apr 8, 2026
Updated 8d ago

Evaluation Results

MethodLinks
2025.09
50.2
2025.09
50.1
2026.03
48.8
2025.09
47.7
2026.01
47.68
2026.01
45.82
2026.03
45.1
2025.09
44.6
2025.09
44.6
2026.01
44.36
2026.03
44.3
2026.01
44.11
2026.01
43.56
2026.01
42.63
2025.09
42.5
2025.09
42.4
2026.01
42.34
2026.01
41.71
2026.01
41.68
2026.01
41.39
2026.02
41.2
2026.04
40.9
2025.12
40.7
2026.01
40.68
2026.02
40.5
2026.03
40.5
2025.09
40.2
2025.09
40
2026.01
39.75
2026.02
39.5
2026.01
39.35
2025.09
39.1
2025.09
39
2026.02
38.8
2026.04
38.8
2025.12
38.2
2026.02
38.1
2026.01
37.39
2026.01
37.39
2026.01
37.31
2026.02
37.2
2026.03
37
2026.01
36.61
2026.04
36.2
2025.09
36
2026.02
35.9
2026.01
35.01
2026.04
34.5
2026.02
33.2
2026.04
33.1
2026.04
32.9
2026.02
32.6
2026.02
32.5
2026.02
32.4
2026.03
32.4
2026.02
31
2026.04
30.7
2025.12
30.5
2026.02
30.3
2026.04
30.2
2026.03
29.9
2026.04
29.8
2025.12
29.7
2026.04
29.7
2025.09
29.5
2026.02
28.9
2026.02
28.9
2026.02
28.1
2025.09
28.1
2025.09
27.9
2026.04
27.4
2025.09
26.8
2026.02
26.7
2026.02
25.5
2025.12
25.5
2026.03
25.5
2026.04
25.5
2026.02
24.1
2025.12
22.1
2026.03
22.1
2025.09
22.1
2026.04
22.1
2025.09
21.7
2025.12
20.8
2025.12
20.1
2026.04
20.1
2025.09
19.4
2025.09
19.2
2026.03
18.7
2025.09
18.7
2025.12
18.6
2026.04
18.6
2026.02
18.3
2026.03
18.3
2025.09
18.3
2026.02
16.4
2025.12
16.4
2026.03
16.4
2025.09
16.2
2025.09
15.8
Showing 100 of 117 rows