Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Safety Alignment Evaluation on OOD Safety Suite Average
Loading...
0.5
Average Absolute Improvement
Sqrt-Competence
-30.076
-22.138
-14.2
-6.262
May 25, 2026
Average Absolute Improvement
Updated 7d ago
Evaluation Results
Method
Method
Links
Average Absolute Improvement
Sqrt-Competence
Backbone=Qwen3-8B
2026.05
0.5
Curri-DPO
Backbone=Yi-1.5-9B
2026.05
-4.3
Sequential
Backbone=Yi-1.5-9B
2026.05
-5.5
Sqrt-Competence
Backbone=Yi-1.5-9B
2026.05
-5.6
Curri-DPO
Backbone=LLaMA-3-8B
2026.05
-6.4
Staged-Competence
Backbone=Yi-1.5-9B
2026.05
-7.1
Sequential
Backbone=Qwen3-8B
2026.05
-7.2
Sqrt-Competence
Backbone=LLaMA-3-8B
2026.05
-9
Curri-DPO
Backbone=Qwen3-8B
2026.05
-10
Sequential
Backbone=LLaMA-3-8B
2026.05
-11.4
Staged-Competence
Backbone=LLaMA-3-8B
2026.05
-12.2
Staged-Competence
Backbone=Qwen3-8B
2026.05
-28.9
Feedback
Search any
task
Search any
task