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

Algebraic Reasoning on AQuA-RAT

66.8Accuracy

ReFT

22.91234.30645.757.094Feb 28, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.02
66.82
2026.02
66.71.9
2026.02
650.2
2026.02
64.80
2026.02
47.70
2026.02
47.60.1
2026.02
45.4-2.3
2026.02
45.4-2.3
2026.02
36.50.3
2026.02
36.20
2026.02
33.6-2.9
2026.02
32.70
2026.02
32.1-0.6
2026.02
31.6-1.1
2026.02
30.2-6
2026.02
24.6-8.1