Recurrent Looped Transformer gives each token 96 logical blocks of depth

The Recurrent Looped Transformer (RLT) pairs a causal encoder with a recurrent decoder that carries its final hidden state and sliding-window attention cache across every prompt and response token. This design enables latent reasoning with unbounded temporal depth, where the path through decoder blocks grows with sequence length while per-token block count stays fixed. It unifies pretraining, SFT, sampling, and RL replay under one complete-state transition, though realized reasoning gains and hardware efficiency remain to be established.