LoopFormer is a Transformer that thinks in loops and can flex its thinking time up or down based on the compute you give it.
Chain-of-Thought (CoT) makes AI think step by step, but it is slow because it writes many tokens one by one.
Render-of-Thought (RoT) turns the modelβs step-by-step thinking from long text into slim images so the model can think faster with fewer tokens.