On the SPOT: Enhancing Speculative Decoding for Diffusion Language Models via Commit-Order Prediction

Fri, 30 Oct 2026 00:00:00 +0900

Paper

Diffusion language models must decide ‘what’ and ‘where’ at the same time

Diffusion language models generate sentences by repeatedly filling masked tokens. To decode exactly, each model forward must commit only the single most certain masked position as a token, which is slow; committing multiple positions at once to increase speed degrades generation quality. This paper moves speculative decoding, a lossless acceleration technique validated on autoregressive models, to diffusion language models, but finds the solution by predicting not tokens but the order of which position to fill next.

Diffusion Language Models on Jaehun's Blog

On the SPOT: Enhancing Speculative Decoding for Diffusion Language Models via Commit-Order Prediction

Diffusion language models must decide ‘what’ and ‘where’ at the same time