Recurrent Looped Transformer hits 341 GitHub stars in under a day for a report that states it has no measured results

GitHub·medium signalXBlueskyLinkedInCopy link

yifanzhang-pro/recurrent-looped-tranformer was created 2026-09-12T19:02Z and reached 341 stars and 45 forks by the following midday. The report pairs a causal encoder with a recurrent decoder that carries its final hidden state and layerwise sliding-window attention cache across every prompt and response token, in a 48-encoder-layer, 48-decoder-layer configuration with attention and FFN weights shared across stages. The project page says outright that it is a specification rather than an empirical report and that "realized reasoning gains, hardware efficiency, and RL scaling remain to be established"; there is no arXiv version.