Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Model Stealing Defense on MedQA
Loading...
62.3
Agreement
No Defense
53.772
55.986
58.2
60.414
Jun 14, 2026
Agreement
Average Value
User Impact
Updated 1mo ago
Evaluation Results
Method
Method
Links
Agreement
Average Value
User Impact
No Defense
Query budget (B)=500,...
2026.06
62.3
44
0
PRADA
Query budget (B)=500,...
2026.06
61.8
43.6
0.4
DRW
Query budget (B)=500,...
2026.06
60.4
43.1
4.8
HoneypotNet
Query budget (B)=500,...
2026.06
59.7
42.4
1.5
KT w/o Breadcrumb
Query budget (B)=500,...
2026.06
59.1
41.4
0
KT w/o Detection
Query budget (B)=500,...
2026.06
58.4
40.3
0
Output Pert.
Query budget (B)=500,...
2026.06
58.1
42.6
2.1
KT (Full)†
Query budget (B)=500,...
2026.06
54.1
37.8
0
Feedback
Search any
task
Search any
task