Gumbel Straight Flow: Distilling Autoregressive Models into One-step Flow Maps

Excerpt

arXiv:2610.00497v1 Announce Type: new Abstract: We present Gumbel Straight Flow (GSF), a continuous flow map language model that leverages the noise-data coupling of a pretrained autoregressive language (AR) model. We theoretically demonstrate that the coupling between Gumbel noise and one-hot token sequences induced by an autoregressive model yields non-intersecting linear paths connecting the noise to the sequence representations. To further enhance high-quality few-step path sampling, we use