Astra’s hidden computation looks more like belief propagation than CoT
Astra solves randomized Boolean-logic problems without visible chain-of-thought, then gets markedly better when given meaningless filler tokens to process. Tracking its per-variable confidence suggests the scores oscillate before converging toward posterior probabilities, much like iterative belief propagation. If the result holds up, “no CoT” does not mean “no multi-step computation”: useful reasoning may be happening in a hidden continuous state.