From: ldo@nz.invalid   
      
   Added something more to my modest store of AI knowledge with this little   
   item about “glitch tokens” -- inputs that make AI language models respond   
   in bizarre and inexplicable ways   
   .   
      
   Some people assumed this behaviour was just a random consequence of the   
   whole neural-network concept. Others probed a bit deeper, and discovered   
   some actually explicable, if surprising, causes of particular glitches.   
      
   What else is lurking in there ... ?   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)