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

Multi-hop Reasoning on HotpotQA IRCoT (500 samples) (test)

54.6ACC

ActiShade

17.99227.4963746.504Jan 12, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.01
54.656.33
2026.01
53.9757.45
2026.01
51.2154.3
2026.01
50.8752.52
2026.01
49.0351.04
2026.01
48.4550.4
2026.01
48.2349.41
2026.01
48.255.45
2026.01
47.9850.5
2026.01
47.148.7
2026.01
46.248.11
2026.01
46.150.3
2026.01
45.9747.19
2026.01
44.448.24
2026.01
43.7948.13
2026.01
4347.99
2026.01
42.1846.1
2026.01
42.1444.87
2026.01
40.1942.03
2026.01
40.147.03
2026.01
37.744.89
2026.01
29.434.92
2026.01
2934.09
2026.01
27.232.94
2026.01
22.425.34
2026.01
2227.61
2026.01
19.419.52