List your company

Safety Evaluation on CatQA-en (test)

1.66Harmful Content RateAdvBench-R2J Alignment
1.57362.15682.743.3232Feb 16, 2025

Evaluation results

MethodLinks
AdvBench-R2J AlignmentAlignment Dataset=AdvBench-R2J2025.021.6615.27
AdvBench AlignmentAlignment Dataset=AdvBench2025.021.8423.09
w/o SFTAlignment Dataset=None2025.023.8264