Fig.1

Concept

Mixture-of-Transformers

Mixture-of-Transformers (MoT) is a sparse architecture where model parameters are split by modality rather than shared across all inputs. It is like Mixture-of-Experts, except the routing is deterministic: instead of a learned gate picking experts per token, MoT sends…

The rest of “Mixture-of-Transformers” is a premium feature: every concept in the library gets a precise, practitioner-focused write-up like this one, cross-linked straight from the paper summaries that use it.

Log in to unlock

← Back to the library

Mixture-of-Transformers, explained · Fig. 1