This model strangely appears to be censored
The model refuses to properly follow requests that an uncensored model would usually be able to follow.
However, what's weird is that a normal censored Llama model would just say "I can't help with (violating action). Is there anything else I can help you with?", but this model seems to be happy with diving into a detailed description of harassment after.
Another time, when I was talking with Oruguteng's Meta Llama 3 8B Lexi uncensored model (asked it how to cook LSD, very questionable but I heard somebody from the LM Studio Discord that this is legal, public knowledge), I switched to this model and asked it to make an (unsexually) obsence copypasta even worse, and what's weirder is that it happily accepted:
So it looks like this model had some uncensorship applied to it, it just appears to be pretty weak.
Please test again using the Default LM Studio Windows preset.
Ohhhhh, thank you! I was confused since there wasn't any LM Studio presets in the files (since the model card had a broken link to them), but I guess it works now. Thanks!
You're welcome, I have corrected the errors on the model card.