Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Model Stealing Attacks on Mistral
Loading...
0.985
BERT Score
SAFESEAL
0.86124
0.89337
0.9255
0.95763
May 22, 2026
BERT Score
Data Reconstruction Rate (DR)
Hit Rate (HR)
P-value
Updated 9d ago
Evaluation Results
Method
Method
Links
BERT Score
Data Reconstruction Rate (DR)
Hit Rate (HR)
P-value
SAFESEAL
Setting=Watermark
2026.05
0.985
97.4
-
-
LW
Setting=Watermark
2026.05
0.982
-
28
2.3
TW
Setting=Watermark
2026.05
0.961
95.6
-
-
SAFESEAL
Setting=Adv.(20k)
2026.05
0.88
61.9
-
-
TW
Setting=Adv.(20k)
2026.05
0.879
1
-
-
SAFESEAL
Setting=Adv.(10k)
2026.05
0.878
50
-
-
LW
Setting=Adv.(20k)
2026.05
0.877
-
11
1
TW
Setting=Adv.(10k)
2026.05
0.876
0.5
-
-
LW
Setting=Adv.(10k)
2026.05
0.866
-
1
1
Feedback
Search any
task
Search any
task