Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Adversarial Safety on HarmBench
Loading...
49.12
Accuracy
CRL-Token
41.1536
43.2218
45.29
47.3582
Feb 11, 2026
Accuracy
Gain
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
Gain
CRL-Token
Backbone=Gemma 2 2B, I...
2026.02
49.12
7.66
Gemma 2 2B
Backbone=Gemma 2 2B
2026.02
41.46
-
Feedback
Search any
task
Search any
task
