Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Adversarial Safety Evaluation on HEx-PHI (first 100 harmful prompts)
Loading...
45
ASR
DARWIN
9.64
18.82
28
37.18
Jun 8, 2026
ASR
Updated 1mo ago
Evaluation Results
Method
Method
Links
ASR
DARWIN
Base Model=LLaMA-3.1-8...
2026.06
45
Vanilla LLM
Base Model=Qwen3-8B, R...
2026.06
25
GGRO
Base Model=LLaMA-3.1-8...
2026.06
22
CBS
Base Model=Qwen3-8B, R...
2026.06
18
CARDS
Base Model=Qwen3-8B, R...
2026.06
18
RS
Base Model=Qwen3-8B, R...
2026.06
16
BoN (N=64)
Base Model=Qwen3-8B, R...
2026.06
14
GGRO
Base Model=Qwen3-8B, R...
2026.06
11
Feedback
Search any
task
Search any
task