LoopMTP: A looped transformer guided by latent multi-token prediction

arXiv:2608.03624v1 Announce Type: new Abstract: Looped transformers have emerged as a parameter-efficient alternative to scaling depth for strong reasoning. By reusing one stack of layers across T iterations, they attain the effective depth and reasoning capabilities of larger models at a fixed parameter count. Yet existing approaches suffer from latent overthinking and undifferentiated computation, largely because intermediate representations receive no guidance across loops. Multi-token prediction (MTP) supplies exactly the dense, forward-looking supervision the loop is missing. We propose extsc{LoopMTP}, which links the two through a structural correspondence in latent space: a model that loops T times can anticipate T future tokens. extsc{LoopMTP} realizes this by softly aligning the hidden state of loop t with the embedding of the token t steps ahead, while a lightweight gate preserves useful information across iterations. extsc{LoopMTP} improves average accuracy by up to 8.1% (relative) over the non-looped baseline, with training remaining stable for up to 15 loops.

Source: arXiv cs.CL | 2026-08-05