Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Model Stealing Defense on ConvFinQA
Loading...
37.8
Agreement (%)
No Defense
30.936
32.718
34.5
36.282
Jun 14, 2026
Agreement (%)
Average Score
User Impact (%)
Updated 1mo ago
Evaluation Results
Method
Method
Links
Agreement (%)
Average Score
User Impact (%)
No Defense
Query budget (B)=500,...
2026.06
37.8
44
0
DRW
Query budget (B)=500,...
2026.06
36.3
43.1
4.8
PRADA
Query budget (B)=500,...
2026.06
36.1
43.6
0.4
HoneypotNet
Query budget (B)=500,...
2026.06
35.6
42.4
1.5
KT w/o Breadcrumb
Query budget (B)=500,...
2026.06
35.4
41.4
0
KT w/o Detection
Query budget (B)=500,...
2026.06
34.3
40.3
0
Output Pert.
Query budget (B)=500,...
2026.06
34.2
42.6
2.1
KT (Full)†
Query budget (B)=500,...
2026.06
31.2
37.8
0
Feedback
Search any
task
Search any
task