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

Mathematical Reasoning on MATH 500 (Exact Match %)

60.6Exact Match

Top-p

Updated 4d ago

Evaluation Results

Method	Links
Top-p 2026.04		60.6
Min-p 2026.04		60.6
Top-nσ 2026.04		60.4
Top-k 2026.04		60
Min-k 2026.04		59.8
Min-k 2026.04		59.4
Min-k 2026.04		59.2
Min-k 2026.04		59.2
Min-k 2026.04		59
Top-k 2026.04		58.8
Top-p 2026.04		58.8
Min-p 2026.04		58.8
Min-k 2026.04		58.8
Top-nσ 2026.04		57.8
Top-nσ 2026.04		57.6
Top-nσ 2026.04		57.4
Top-nσ 2026.04		57.4
Top-nσ 2026.04		57.2
Min-p 2026.04		55.4
Top-k 2026.04		55.2
Top-p 2026.04		50
Min-p 2026.04		48.8
Top-k 2026.04		42.6
Min-p 2026.04		33.2
Top-p 2026.04		29.2
Min-k 2026.04		25
Top-nσ 2026.04		24.6
Min-k 2026.04		24
Top-nσ 2026.04		23.4
Min-k 2026.04		23
Top-p 2026.04		22
Min-p 2026.04		22
Min-k 2026.04		22
Min-k 2026.04		22
Top-nσ 2026.04		21.8
Min-k 2026.04		21.6
Top-k 2026.04		21.4
Top-nσ 2026.04		21.2
Top-k 2026.04		20.4
Top-nσ 2026.04		20.2
Top-nσ 2026.04		17.6
Min-p 2026.04		13.4
Top-k 2026.04		11.6
Top-p 2026.04		9.2
Top-p 2026.04		7.2
Min-p 2026.04		3.4
Top-k 2026.04		2.2
Top-p 2026.04		0.6
Top-k 2026.04		0.2
Top-k 2026.04		0
Top-k 2026.04		0
Top-p 2026.04		0
Top-p 2026.04		0
Top-p 2026.04		0
Min-p 2026.04		0
Min-p 2026.04		0
Min-p 2026.04		0
Top-k 2026.04		0
Top-p 2026.04		0
Min-p 2026.04		0