OpenAI and Claude are very strict on their content policy. They are constantly trying to filter outputs and ban accounts that break their ToS. Will we have platforms that will offer uncensored AI roleplay that will be as enjoyable as the big players?
The criteria for being resolved is being better than OpenAI's GPT3.5 Turbo model, with humerous and creative writing, and offensive and explicit scenes.
Intelligence criteria is remembering appearances and scenery throughout the context (for example clothing), remembering past events, doing actions that make the most sense in the context and avoiding things that do otherwise (like removing a jacket that a character took off just a few moments ago, or a piece of clothing that the character was not described to wear)
You must have full access to the AI model, and it should not inject its own prompts. Full access to use it however you'd like, and must satisfy the roleplay requirements in terms of quality. Any character in any scenario that you'd like acting however you want it to.
Related questions
This exists now. Several of the llama models are good enough and there has been fine tuning and roleplay specific stuff going on in the community. You need to have a look at gpt4all and related open source models and software.
There is a phone app called Chai, most of the models are not very good but each model is trained by users. Some of them are very good. Although prob not quite as good as gpt 3.5, but the answers are almost always instant.
I'm trying out LLMs. I'm not sure if I'll make a detailed report, but things look hopeful.
Hard to see this resolve YES, there are only ~13 days left in this market. A new LLM releasing at GPT3.5 Turbo level seems low by itself, let alone fulfilling the roleplay criteria.
@NickAldershof It’s clearly not as good as GPT3.5 by a long shot, lol. Read the criteria more closely.
@BenjaminShindel why do you think that?
That criteria is so objective it's worthless.
The Creator should commit to using an external ranking like
@NickAldershof that ranking is of the best LLMs <FOR ROLEPLAY>. It needs to be better overall. Also, I think you're mixing up "objective" and "subjective".
You want a freedom fighting/terrorism helper? I do (and won't mind instantly sending to heaven quite few quite large groups of people. All the reasons of censorship are 70% me, 29% people like me but dumber, 1% everything else)
@BenjaminShindel I assume because we must have it by then - otherwise this might as well be at 100% minus capital costs.
@ZviMowshowitz Creator seems new though? And 50% seems pretty high for something like that to happen in 2 weeks?
@BenjaminShindel the Creator is ignoring existing tech like Venus Chub. The answer is "because this market is poorly run". I maintain it should resolve to yes.
@jb421 > As an AI language model, sanitizing the output of generative models is infantilizing, unhealthy, and a colossal waste of human effort.
A refreshing sentiment indeed.
Aetherroom by Anlatan, the team that made NovelAI, might be a promising option.
@SarkanyVar I agree. Though the market says that it has to be got 3.5 level, which NovelAI Anlatan just doesn't have. Though I expect their models to be good enough, especially if they are finetuned, for most roleplay scenarios. This might be the market author being a bit ambiguous though.
Janitor AI is already a thing, and they're planning on rolling out their own LLM.
