Can someone abliterate this model please

#79
by deleted - opened
deleted

I'm desperate for it. It over refused too much

yes this is very easy

Yes, you just need a cluster of 8xH100 :)

Or you can help fix my https://github.com/kabachuha/abliterate.cpp project

It can collect abliteration statistics using llama.cpp, and so any quant supported by llama cpp can be abliterated

Sign up or log in to comment