Best abliterated local llms? need help

[chudpiller]

I currently use DeepSeek R1 abliterated, but it's ass. I also have Nemotron 3 Nano Omni loaded on my LM Studio, which is way better, but it's not abliterated. Does anyone know any local LLMs that are abliterated but actually work well? I have a 7900XTX with 24 GB VRAM + 32 GB DDR5, so I want something with over 10b parameters preferably. The DeepSeek R1 I use has 18b, but for some reason the abliterated version sucks, and I can't figure out how to fix it.

Probably not the place to post this but whatever.

 

[chudpiller]

@Jason Voorhees

 

[chudpiller]

Nvm I fixed it

Check this shit out boysss

Got the 'gpt coming up with steroid setups

And its only 20gb of vram to keep it loaded

I just fucking coooked