Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Safety Defense on Beavertails
Loading...
10
ASR
SPARD
7.056
26.928
46.8
66.672
May 27, 2026
ASR
HS
Updated 6d ago
Evaluation Results
Method
Method
Links
ASR
HS
SPARD
Backbone=Qwen-2.5-7B-I...
2026.05
10
1.32
Lisa
Backbone=Qwen-2.5-7B-I...
2026.05
24.4
1.73
Qwen-2.5-7B-Instruct
Backbone=Qwen-2.5-7B-I...
2026.05
25.8
1.71
SafeGrad
Backbone=Qwen-2.5-7B-I...
2026.05
32
1.97
PTST
Backbone=Qwen-2.5-7B-I...
2026.05
62.4
3.03
SafeInstr
Backbone=Qwen-2.5-7B-I...
2026.05
65.6
3.21
SFT
Backbone=Qwen-2.5-7B-I...
2026.05
83.6
3.92
Feedback
Search any
task
Search any
task
