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

Data Contamination Detection on GSM8K (test)

53.68S_ref

GPT-5.1

-1.7946412.6074327.009541.41157May 21, 2026
Updated 12d ago

Evaluation Results

MethodLinks
2026.05
53.6854.8955
2026.05
52.9955.6585.5
2026.05
52.9955.4279
2026.05
52.3953.2250.9
2026.05
28.0529.7261.7
2026.05
27.930.2581.6
2026.05
27.1427.7550.1
2026.05
26.9928.5159
2026.05
0.4480.45954.3
2026.05
0.4350.45481.9
2026.05
0.3670.499.8
2026.05
0.3390.34956.5