Comment by headPoet

ChatJimmy isn't a model, it's Llama 3.1 8B hardwired into silicon. The point isn't to be a good llm, but to showcase the speedup that's possible