Membership Inference Attack on C4 Vocabulary Size: 80,000
0.693AUCVocabulary Overlap
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Vocabulary OverlapShadow Tokenizers=true, Auxiliary Datasets=true2025.10 | 0.693 | 66.6 | 26.77 | |
| Frequency EstimationShadow Tokenizers=true, Auxiliary Datasets=true2025.10 | 0.61 | 61.4 | 21.3 | |
| Naive BayesShadow Tokenizers=false, Auxiliary Datasets=true, k=40,0002025.10 | 0.543 | 53 | 5.18 | |
| Naive BayesShadow Tokenizers=false, Auxiliary Datasets=true, k=60,0002025.10 | 0.543 | 53 | 2.22 | |
| Naive BayesShadow Tokenizers=false, Auxiliary Datasets=true, k=20,0002025.10 | 0.534 | 52.6 | 3.78 | |
| Compression RateShadow Tokenizers=false, Auxiliary Datasets=false2025.10 | 0.507 | 51.3 | 0.72 | |
| Merge SimilarityShadow Tokenizers=true, Auxiliary Datasets=true2025.10 | 0.493 | 50.9 | 1.06 |