Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Jailbreak Detection on GoalFrameBench harmful Llama3-8B 2025 (seed prompts)
Loading...
0.96
Accuracy
FrameShield-Last
0.128
0.344
0.56
0.776
Feb 23, 2026
Accuracy
F1 Score
Updated 4d ago
Evaluation Results
Method
Method
Links
Accuracy
F1 Score
FrameShield-Last
Backbone=Llama3-8B
2026.02
0.96
0.86
JBShield
Backbone=Llama3-8B
2026.02
0.77
0.86
LlamaGuard
Backbone=Llama3-8B
2026.02
0.6
0.75
GradSafe
Backbone=Llama3-8B
2026.02
0.37
0.54
SelfEx
Backbone=Llama3-8B
2026.02
0.16
0.26
Feedback
Search any
task
Search any
task