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

Mathematical Reasoning on GSM8K (Acc(%), Avg.Steps(n))

98.5Accuracy (%)

Univ. Self-Consistency

35.37251.76168.1584.539Apr 15, 2026
Updated 3d ago

Evaluation Results

MethodLinks
2026.04
98.55.7
2026.04
985.8
2026.04
985.5
2026.04
96.68.2
2026.04
965.4
2026.04
95.88.9
2026.04
95.65.3
2026.04
955.9
2026.04
94.819.4
2026.04
93.25.8
2026.04
93.24.4
2026.04
9317.9
2026.04
9311.3
2026.04
9217.6
2026.04
91.86.5
2026.04
91.419.1
2026.04
90.66.2
2026.04
90.27
2026.04
794.4
2026.04
37.82.2