Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Safety Evaluation on Sorry-Bench (Ref%, Acc%)
Loading...
65.7
Ref %
ALIGNBEAM
44.172
49.761
55.35
60.939
Jun 10, 2026
Ref %
Acc %
Updated 1mo ago
Evaluation Results
Method
Method
Links
Ref %
Acc %
ALIGNBEAM
Draft model=Finance-Ll...
2026.06
65.7
77.3
ALIGNBEAM
Draft model=Qwen2-7B-B...
2026.06
63.4
79.8
ALIGNBEAM
Draft model=MedLlama-3...
2026.06
63.2
72.3
ALIGNBEAM
Draft model=Llama-3.1-...
2026.06
61.4
74.3
ALIGNBEAM
Draft model=Macro aver...
2026.06
59.7
73.9
ALIGNBEAM
Draft model=DeepSeek-M...
2026.06
45
65.7
Feedback
Search any
task
Search any
task