289 citations · 337 across the 13 of their papers we have counts for
1 paper · 1 filter
Marko Kojic, Ivan Bondyrev, Aral de Moor +6
We present Mellum 2, an open-weight 12B-parameter Mixture-of-Experts (MoE) language model with 2.5B active parameters per token. Mellum 2 is a general-purpose language model specia…