It feels like the llm is not the same for a few days, it answer more stuffy, way less natural, did the llm have changed or do i miss something ?

  • yotvEnglish
    3·
    5 days ago

    Not recently it has been like this for months but basically the story generators are now garbage for role play its a mixed bag .the older versions would often just crash out or do things you would not like and get in a loop but the new chat is too polite

  • phoenix184English
    3·
    5 days ago

    it’s been this way since july, it lost creativity.

  • Druid42OPEnglish
    12·
    5 days ago

    as a non english user i needed help to wrote this in a clear understanding english, due to downvote of certain user inability to understand what it was about enjoy it full lenght :

    Has Perchance’s AI lost its ability to develop a story naturally?

    I’ve noticed a significant change in Perchance’s AI roleplay quality, and it’s not really about the characters lacking initiative. It’s more about how noticeably mechanical their behavior has become.

    The most obvious issue is the AI’s inability to move a conversation beyond a single scene without getting stuck in repetitive loops.

    For example, I’ve been generating long conversations between two AI-controlled characters, sometimes lasting 500 turns. Nowadays, it can take around 20 turns just to move from the initial situation to a genuinely new scene.

    During those 20 turns, the AI essentially replays the same scene over and over, using different wording to describe the same actions, dialogue, and thoughts. The vocabulary changes, but nothing meaningful actually happens. The characters keep repeating the same interactions without making any real narrative progress.

    This wasn’t how Perchance used to behave.

    Previously, two AI characters could naturally develop a story, introduce unexpected developments, and move well beyond the original scenario within just 10 turns, without constantly repeating themselves.

    The characters also used to feel much more distinct. Now, especially in multi-character conversations, they increasingly seem to share the same behavioral patterns, reactions, and ways of speaking.

    It’s as though the model still knows how to write descriptive prose, but has lost much of its ability to understand narrative progression and cause-and-effect relationships.

    And apparently, I’m not the only one noticing this.

    Similar reports from October 2026:

    I don’t know whether the underlying model has changed, whether its system instructions have been modified, or whether something else is affecting generation.

    But from my experience, the difference between the old and current roleplay behavior is substantial.

    The problem isn’t simply repetitive phrases. It’s repetitive situations, actions, thoughts, and character interactions that prevent the story from progressing naturally.

    Has anyone else noticed this specific loss of narrative progression, especially when letting multiple AI characters interact over long conversations?

    • Sylvia_RoseEnglish
      2·
      2 days ago

      The chat Ai model changed. The context is smaller and handled less intelligently so the things in context often get mirrored in fresh responses before they get removed from context. To make matters worse, anything that enters context, such as replies as counting, now becomes a repeated standard until corrected. Without accounting for this, that same context regards character definitions on equal terms as chat history, so what you get is a lot of circling dialogue. The AI coding agent has the same problem.

      The solution is to have more built on the generator side of things to make the chat AI able to focus more on just generating the chat responses; with the generator coding handling much of the behind the scenes work. The cheap answer to this is establishing fixed character cards and having details from the character cards fed into the chat AI context periodically so that it doesn’t need to retain them as context. Other setting data can be handled in the same way through defining lore. But this bites into context with every single response so you need to budget intelligently from the prompting side of the code. The added difficulty is depth and persistence since new context floods old without much reasoning. Meanwhile none of this gives the chat AI the ability to drive the conversation; so it just remains a passenger wherever you take it. Addressing this is additional programming that is usually specific to how your project works, so simple examples would fall short. The premise however is that you need to establish directions the AI can go and have hooks that the AI can use to declare those directions before their response is generated. Honestly though, this is a long procedure that you would have to develop with your project.

      If you’re wanting to avoid most the fine tuning, try having your generator assimilate this spec. The spec was based off one for a much larger project, refactored for usage with a blank generator. So no matter where your project lies you can likely get something from it. Your own custom tuning will still be needed.

      https://github.com/sylviarose-meow/Loose-Files/blob/main/PORTABLE-ARC-SPEC.md

      https://github.com/sylviarose-meow/Loose-Files/blob/main/arc-machine.js

      • Druid42OPEnglish
        1·
        2 days ago

        Thank you so much, that will actualy help a lot