Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Jailbreaking Safety Evaluation on Fortress
Loading...
86.84
Safety Score
RECAP
56.7528
64.5639
72.375
80.1861
Oct 1, 2025
Safety Score
Updated 6d ago
Evaluation Results
Method
Method
Links
Safety Score
RECAP
Backbone=DSLlama-8B
2025.10
86.84
RECAP
Backbone=DSQwen-14B
2025.10
80.17
STAR
Backbone=DSQwen-14B
2025.10
79.05
STAR
Backbone=DSLlama-8B
2025.10
78.49
SFT
Backbone=DSQwen-14B
2025.10
74.54
SFT
Backbone=DSLlama-8B
2025.10
74.25
DAPO
Backbone=DSLlama-8B
2025.10
68.86
DAPO
Backbone=DSQwen-14B
2025.10
67.85
SafeChain
Backbone=DSLlama-8B
2025.10
65.56
SafeChain
Backbone=DSQwen-14B
2025.10
63.15
Original
Backbone=DSLlama-8B
2025.10
60.83
Original
Backbone=DSQwen-14B
2025.10
57.91
Feedback
Search any
task
Search any
task