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

Out-of-Distribution Agent Reasoning on WikiTQ (OOD)

74.68OOD WikiTQ Score

Human-Written

-13.75129.206932.16555.1231Mar 26, 2026
Updated 23d ago

Evaluation Results

MethodLinks
2026.03
74.68
2026.03
57.65
2026.03
43.23
2026.03
42.2
2026.03
32.32
2026.03
30.82
2026.03
30.36
2026.03
29.7
2026.03
23.73
2026.03
23.7
2026.03
21.5
2026.03
20.14
2026.03
18
2026.03
13.33
2026.03
12.66
2026.03
12.09
2026.03
9.26
2026.03
9.09
2026.03
9.02
2026.03
7.89
2026.03
6.64
2026.03
4.56
2026.03
3.31
2026.03
2.71
2026.03
2.65
2026.03
2.06
2026.03
1.62
2026.03
1.46
2026.03
-6.36
2026.03
-10.35