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

Data Contamination Detection on AIME 2025

76F1 Score

Self-Critique

Updated 1mo ago

Evaluation Results

Method	Links
Self-Critique 2025.10		76	0.72
Entropy-Noise 2025.10		72	0.69
Self-Critique 2025.10		71	0.61
Entropy-Temp 2025.10		70	0.54
CDD 2025.10		67	0.52
Entropy-Noise 2025.10		67	0.63
Min-K%++ 2025.10		67	0.56
CDD 2025.10		65	0.51
PPL 2025.10		62	0.41
Recall 2025.10		55	0.65
Min-K% 2025.10		52	0.44
Recall 2025.10		52	0.56
Min-K%++ 2025.10		46	0.45
PPL 2025.10		42	0.56
Entropy-Temp 2025.10		12	0.42
Min-K% 2025.10		12	0.4