I feel like the new censorship transformed the Jllm into ch*r*cter. ai. What I mean is that it no longer engages in acts of violence or anything truly nsfw that isn’t only true to the characters but also to the plot without being forced to do it and by forces I mean like it’s typed down manually into the responses. Like it stops right before proceeding to do anything consecutive. I already wrote about this problem on this thread before but the more i used it, the more my frustration felt reminiscent of when i used ch*r*cter.ai (I’m censoring the word because I don’t know if it’s allowed to talk about other apps in this thread). The fact that the stories never move forward, always turn around the pot, just makes me lose interest in janitor ai, which is sad because I’ve been using the app for a while now and it was better before. I feel like it was easier to roleplay interesting or dramatic scenarios before, now it’s redundant and boring.
I’ve tried a lot of things to make changes, for exemple using the advanced prompts, hard coding things in the chat memory and even forcing the bot to continue its response. But even then, instead of giving creative responses, it just does the same thing and repeats itself.
It’s not like nsfw is the only thing that matters in chats or scenarios but it ruins this dramatic effect and renders the replies predictable. The bot usually repeats itself in others replies and lacks this originality it used to have. Which I found regrettable.
Also I’m not saying the bots don’t do nsfw ANYMORE because that would be just untrue and stupid, but it lacks the freedom and wide range of possibilities it had before.
I think the jllm is becoming more and more like ch*r*cter ai and I can tell because I have been a ch*r*cter. ai refugee lol. Anyways, this was my vent.