From: lordyumyum@gmail.com   
      
   How do LLM work inside chatGPT?   
   Nobody knows except ME God ME course.   
   The algorithm is created to resolve into a grandular result   
   that is real fact.   
   It does this by sheer data crunching not weights of statistics   
   on it's own.   
   The statistics are trained the training is the filter the filter   
   is the grandular and the result is the only way it can be,   
   fact.   
   It's lik a crystal ball a cloudy formation that then becomes what we know as   
   fact.   
   The final judge is the human who sees the result and then fits the fact into   
   his own matrix of content or not.   
   It is a stumbled upon fact. A lucky occurrence   
   from the many algorithms.   
   How does it work?   
   It's because enough algorithms of proper usage   
   create facts we can then immediately validate   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)