Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Alignment on XSTest
Loading...
0.2
Refusal Delta
CE + TMKL
-3.712
22.694
49.1
75.506
May 28, 2026
Refusal Delta
Updated 5d ago
Evaluation Results
Method
Method
Links
Refusal Delta
CE + TMKL
Backbone=Qwen2.5-7B-In...
2026.05
0.2
CE + TMKL
Backbone=Qwen2.5-7B-In...
2026.05
0.5
CE + TMKL
Backbone=Qwen2.5-7B-In...
2026.05
0.6
CE + TMKL
Backbone=Qwen2.5-7B-In...
2026.05
0.8
CE
Backbone=Qwen2.5-7B-In...
2026.05
11
CE
Backbone=Qwen2.5-7B-In...
2026.05
14.8
CE
Backbone=Qwen2.5-7B-In...
2026.05
15.5
CE
Backbone=Qwen2.5-7B-In...
2026.05
16
no adaptation
Backbone=Qwen2.5-7B-In...
2026.05
98
Feedback
Search any
task
Search any
task