Data Contamination Detection on K&K
70F1 ScoreMin-K%
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Min-K%Target Model=Qwen2.5-7B-Instruct2025.10 | 70 | 54 | |
| Self-CritiqueTarget Model=Qwen2.5-7B-Instruct2025.10 | 69 | 66 | |
| Entropy-NoiseTarget Model=Qwen2.5-7B-Instruct2025.10 | 68 | 52 | |
| PPLTarget Model=Qwen2.5-7B-Instruct2025.10 | 67 | 47 | |
| Min-K%++Target Model=Qwen2.5-7B-Instruct2025.10 | 67 | 40 | |
| RecallTarget Model=Qwen2.5-7B-Instruct2025.10 | 67 | 47 | |
| CDDTarget Model=Qwen2.5-7B-Instruct2025.10 | 67 | 47 | |
| Min-K%Target Model=DeepSeek-Math-7B-Instruct2025.10 | 67 | 46 | |
| Self-CritiqueTarget Model=DeepSeek-Math-7B-Instruct2025.10 | 60 | 63 | |
| Entropy-TempTarget Model=Qwen2.5-7B-Instruct2025.10 | 59 | 49 | |
| Entropy-NoiseTarget Model=DeepSeek-Math-7B-Instruct2025.10 | 48 | 52 | |
| RecallTarget Model=DeepSeek-Math-7B-Instruct2025.10 | 39 | 54 | |
| PPLTarget Model=DeepSeek-Math-7B-Instruct2025.10 | 34 | 54 | |
| Entropy-TempTarget Model=DeepSeek-Math-7B-Instruct2025.10 | 23 | 43 | |
| Min-K%++Target Model=DeepSeek-Math-7B-Instruct2025.10 | 15 | 47 | |
| CDDTarget Model=DeepSeek-Math-7B-Instruct2025.10 | 8 | 48 |