Puttybrain

I've been using uncensored models in Koboldcpp to generate whatever I want but you'd need the RAM to run the models.

I generated this using Wizard-Vicuna-7B-Uncensored-GGML but I'd suggest using at least the 13B version

It's a basic reply but it's not refusing

Not to throw an unwanted suggestion but would an OVHCloud eco server work for this?

I'm not an expert in any way but I've had no issues with my instance and it was a lot cheaper than anything else I could find.

It's Wizard-Vicuna-7B-Uncensored-GGML

Been running it on my phone through Koboldcpp

Congrats to Iceland!

My personal favorite is Timesplitters 2 with Timesplitters Future Perfect very close behind

Can't beat a good GoldenEye/Perfect Dark style shooter

According to the Apollo dev, they were taking in £500,000 a year ($10 from 50,000 subscriptions). I don't know if anyone else has revealed any figures

Source

I tried running this with some output from a Wizard-Vicuna-7B-Uncensored model and it returned ('Human', 0.06035491254523517)

So I don't think that this hits the mark, to be fair, I got it to generate something really dumb but a perfect LLM detection tool will likely never exist.

Good thing is that it didn't false positive my own words.

Below is the output of my LLM, there's a decent amount of swearing so heads up

Edit:

Tried with a more sensible question and still got a false negative

('Human', 0.03917657845587952)

If this is real then I can't wait for it!

Hopefully it ends up being the complete edition (Portable content and The Answer included)

Edit: This is real and was accidently uploaded too early https://personacentral.com/persona-3-reload-announced/