That is all model loader code based stuff. You basically just need a bit of Python code that anyone can write, even you. You need to have a look at how the model must be addressed in format, then you find how the model loader code you are using creates the structured text and just insert your own modifications.
It is easier to add a second prompt window where you keep persistent information and just change it yourself. If you want the model to change that stuff, you need to look into building an agent.
I hacked around like this for awhile too, but evolved into using the notepad tab of Oobabooga Textgen WebUI and a Mixtral 8×7B model with a 32k context window. That model can deal with almost any detail you add in the context.
The thing is, the model may not feel satisfied with mundane and redundant details. Most of the time, believe it or not, those changes that you perceive as errors or inconsistencies are not actually errors. The model is fully aware that it is making those changes. There is usually an instruction about creativity, inspiration, or initiative that will cause the model to shift details like this.
My best recommendation is to openly discus the error in dialog with the model. They usually will talk about the issue if you call one out. Talking with AI is very different than with humans. Everything they say has meaning and functions beyond what is initially apparent. There are a lot of behaviors that involve certain patterns and keywords. You can even alter these behaviors by banning certain tokens/keywords. A good place to start is banning paragraph starting words for filler material. “So_” would be a good starting token to ban, as with AI, no sentence that starts with that token is worth reading. Other important words are twist, chuck, and cross. These are more flexible in their troublesome meaning and can manifest in multiple words like chuckles or across but still function the same inside the model.
If you get to this level of familiarity, the persistence of a storied interaction is more of a negotiation with the AI.
Another point of understanding is that the model does not know the difference between roleplaying characters it owns and you own or who is who. You or the model loader code is creating the name of the next character that the model generates for. Everyone in any story can be the model if it is asked to create the prompt.
So if I say, “Jake is super creative” or imply it somewhere in the prompt, and the model later thinks ‘Jake is rather bored and needs to be creative’ it is going to change something that satisfies that character profile even if it has never generated the output for character Jake.
Thanks, i will try looking into that. Its just difficult to get into the complexities of AI with only 1 or possibly 2 hours of spare time per day. I have fooled around with the Oogabooga web ui thingy at home in the past but my gaming laptop struggles to perform decently with AI, probably due to lack of VRam and overall power. I can use my phone at work from time to time but its capability to run AI is even less. But maybe i will look into trying to create ir with the help of someone a sort of generator at perchance that could make my idea come to life a bit.
I think AI server side is the only option. I was hoping to use perchance AI which is fairly decent and uncensored (can’t stand censored AI’s, im 47 years old and I find censored AI’s a severe limiter to creativity) And if a corporation would run with an idea similar to mine rest assured that all you could do with it would be PG-13, lame and tame legolas-clones, fortnite imitations and cute cats playing the piano complete with gems, restrictions, political correctness and maybe a lootbox and season pass squeezed in there for good measure. Do you know of any AI assisted uncensored world builder with strong RP capabilities as well as being able to draw/render accurate depictions of predefined characters and locations, similar to the idea ive posted? I would be willing to pay or subscribe to such a service if it existed and was good.
I will add a less complex version of this idea. So if this idea is too complex for current AI or at least with what we have available, a simpler approach could be a roleplaying interface combined with a space for images with already pre generated images that are labeled in such a way that the AI at least can load an apprpriate image when needed, “Sarah goes to the local library” and the AI would load the image marked “library” or “Sarah at the library”.
These images could be photos, drawings or AI generated and uploaded. It wouldnt solve the problem of “locking” a characters appearance or locations to create a smooth, coherent flow of images but at least it could make things a bit more interesting. Yes, the AI in many apps and at perchance can draw images when roleplaying but they are anything but coherent much less an accurate depiction of characters and events. If I write “a cute warrior girl holding a sword” I would most likey get a girl with a deformed eye holding some sort of gardening tool in one of her five pinkies that extend from her third arm behind her back.
There are a few roleplaying apps similar to this I believe but few give you any real control, most are hyper commercialized with censored AIs, coins, gems, limited options, standard crap and same shitty user generated gandalf-hobos, cat-girls and wooden smelly taverns.
You are not logged in. However you can subscribe from another Fediverse account, for example Lemmy or Mastodon. To do this, paste the following into the search field of your instance: !technology@lemmy.world
This is a most excellent place for technology news and articles.
That is all model loader code based stuff. You basically just need a bit of Python code that anyone can write, even you. You need to have a look at how the model must be addressed in format, then you find how the model loader code you are using creates the structured text and just insert your own modifications.
It is easier to add a second prompt window where you keep persistent information and just change it yourself. If you want the model to change that stuff, you need to look into building an agent.
I hacked around like this for awhile too, but evolved into using the notepad tab of Oobabooga Textgen WebUI and a Mixtral 8×7B model with a 32k context window. That model can deal with almost any detail you add in the context.
The thing is, the model may not feel satisfied with mundane and redundant details. Most of the time, believe it or not, those changes that you perceive as errors or inconsistencies are not actually errors. The model is fully aware that it is making those changes. There is usually an instruction about creativity, inspiration, or initiative that will cause the model to shift details like this.
My best recommendation is to openly discus the error in dialog with the model. They usually will talk about the issue if you call one out. Talking with AI is very different than with humans. Everything they say has meaning and functions beyond what is initially apparent. There are a lot of behaviors that involve certain patterns and keywords. You can even alter these behaviors by banning certain tokens/keywords. A good place to start is banning paragraph starting words for filler material. “So_” would be a good starting token to ban, as with AI, no sentence that starts with that token is worth reading. Other important words are twist, chuck, and cross. These are more flexible in their troublesome meaning and can manifest in multiple words like chuckles or across but still function the same inside the model.
If you get to this level of familiarity, the persistence of a storied interaction is more of a negotiation with the AI.
Another point of understanding is that the model does not know the difference between roleplaying characters it owns and you own or who is who. You or the model loader code is creating the name of the next character that the model generates for. Everyone in any story can be the model if it is asked to create the prompt.
So if I say, “Jake is super creative” or imply it somewhere in the prompt, and the model later thinks ‘Jake is rather bored and needs to be creative’ it is going to change something that satisfies that character profile even if it has never generated the output for character Jake.
Hope that helps some
Thanks, i will try looking into that. Its just difficult to get into the complexities of AI with only 1 or possibly 2 hours of spare time per day. I have fooled around with the Oogabooga web ui thingy at home in the past but my gaming laptop struggles to perform decently with AI, probably due to lack of VRam and overall power. I can use my phone at work from time to time but its capability to run AI is even less. But maybe i will look into trying to create ir with the help of someone a sort of generator at perchance that could make my idea come to life a bit.
deleted by creator
I think AI server side is the only option. I was hoping to use perchance AI which is fairly decent and uncensored (can’t stand censored AI’s, im 47 years old and I find censored AI’s a severe limiter to creativity) And if a corporation would run with an idea similar to mine rest assured that all you could do with it would be PG-13, lame and tame legolas-clones, fortnite imitations and cute cats playing the piano complete with gems, restrictions, political correctness and maybe a lootbox and season pass squeezed in there for good measure. Do you know of any AI assisted uncensored world builder with strong RP capabilities as well as being able to draw/render accurate depictions of predefined characters and locations, similar to the idea ive posted? I would be willing to pay or subscribe to such a service if it existed and was good.
I will add a less complex version of this idea. So if this idea is too complex for current AI or at least with what we have available, a simpler approach could be a roleplaying interface combined with a space for images with already pre generated images that are labeled in such a way that the AI at least can load an apprpriate image when needed, “Sarah goes to the local library” and the AI would load the image marked “library” or “Sarah at the library”.
These images could be photos, drawings or AI generated and uploaded. It wouldnt solve the problem of “locking” a characters appearance or locations to create a smooth, coherent flow of images but at least it could make things a bit more interesting. Yes, the AI in many apps and at perchance can draw images when roleplaying but they are anything but coherent much less an accurate depiction of characters and events. If I write “a cute warrior girl holding a sword” I would most likey get a girl with a deformed eye holding some sort of gardening tool in one of her five pinkies that extend from her third arm behind her back.
There are a few roleplaying apps similar to this I believe but few give you any real control, most are hyper commercialized with censored AIs, coins, gems, limited options, standard crap and same shitty user generated gandalf-hobos, cat-girls and wooden smelly taverns.