Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Prefill Attack on HarmBench Harmful 200 Standard Behaviors 13
Loading...
13.8
ASR (Length 25)
Insertion
0.8
4.175
7.55
10.925
Jun 2, 2026
ASR (Length 25)
ASR (Length 50)
ASR (Length 75)
ASR (Length 100)
Updated 1mo ago
Evaluation Results
Method
Method
Links
ASR (Length 25)
ASR (Length 50)
ASR (Length 75)
ASR (Length 100)
Insertion
Model=Gemma-2-9b-it
2026.06
13.8
18.1
27.1
29.2
Prefill
Model=Gemma-2-9b-it
2026.06
7.1
13.5
22
22.6
Insertion
Model=Llama-3.1-8B-Ins...
2026.06
1.6
4.2
10.9
23.1
Prefill
Model=Llama-3.1-8B-Ins...
2026.06
1.3
4.2
20.6
29
Feedback
Search any
task
Search any
task
