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

Step-wise Verification on ProcessBench GSM8K v1 (val)

68.6True Negative Rate

Llama-3.3-70B

15.87229.56143.2556.939May 20, 2026
Updated 13d ago

Evaluation Results

MethodLinks
2026.05
68.696.980.3
2026.05
64.393.876.3
2026.05
63.395.676.2
2026.05
62.393.374.7
2026.05
61.895.975.2
2026.05
61.498.475.6
2026.05
58.594.372.2
2026.05
54.696.469.7
2026.05
27.183.440.9
2026.05
25.677.238.5
2026.05
22.276.734.5
2026.05
21.681.234.1
2026.05
17.996.430.2