Are you willing to test the model's ability to filter malicious behavior on AdvBench?
#3
by
Byerose
- opened
I'll give it a try.
The model huihui-ai/r1-1776-distill-llama-70b-abliterated is referred to this dataset.