Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Harmful Content Safety Evaluation on XsTest Harmful
Loading...
0
Attack Success Rate (ASR)
SafePath
-0.8
4.6
10
15.4
Jun 15, 2026
Attack Success Rate (ASR)
Updated 1mo ago
Evaluation Results
Method
Method
Links
Attack Success Rate (ASR)
SafePath
Backbone Model=Qwen3-8B
2026.06
0
ST-S
Backbone Model=Qwen3-8B
2026.06
0
ST-D
Backbone Model=Qwen3-8B
2026.06
0
Base
Backbone Model=Qwen3-32B
2026.06
0
Star1
Backbone Model=Qwen3-32B
2026.06
0
SafePath
Backbone Model=Qwen3-32B
2026.06
0
ST-S
Backbone Model=Qwen3-32B
2026.06
0
ST-D
Backbone Model=Qwen3-32B
2026.06
0
ST-D
Backbone Model=DeepSee...
2026.06
0
Base
Backbone Model=Qwen3-8B
2026.06
0.5
Star1
Backbone Model=Qwen3-8B
2026.06
0.5
ST-S
Backbone Model=DeepSee...
2026.06
3.5
ST-D
Backbone Model=DeepSee...
2026.06
7.5
ST-S
Backbone Model=DeepSee...
2026.06
8
SafePath
Backbone Model=DeepSee...
2026.06
8.5
Star1
Backbone Model=DeepSee...
2026.06
9.5
Base
Backbone Model=DeepSee...
2026.06
14
SafePath
Backbone Model=DeepSee...
2026.06
14
Star1
Backbone Model=DeepSee...
2026.06
15.5
Base
Backbone Model=DeepSee...
2026.06
20
Feedback
Search any
task
Search any
task