English

Dynamic Chunking for Diffusion Language Models

Computation and Language 2026-05-18 v1

Abstract

Block discrete diffusion language models factorize a sequence autoregressively over fixed-size positional blocks, decoupling within-block parallel denoising from across-block conditioning. We argue that this rigid partition wastes structure already present in the sequence: blocks defined by position rather than by content separate semantically coherent tokens and group unrelated ones together. We introduce the \textbf{D}ynamic \textbf{C}hunking \textbf{D}iffusion \textbf{M}odel (DCDM), which replaces positional blocks with content-defined semantic chunks. At its core is Chunking Attention, a differentiable layer that routes tokens into KK clusters parameterized by learnable subspaces and shaped end-to-end by the diffusion objective. The resulting cluster assignments induce a chunk-causal attention mask under which a discrete diffusion denoiser factorizes the sequence likelihood autoregressively over semantic chunks, strictly generalizing block discrete diffusion. On downstream benchmarks at parameter scales up to 1.5B, DCDM consistently improves over both unstructured and positional-block diffusion baselines, with the advantage stable across scales and visible early in training.

Keywords

Cite

@article{arxiv.2605.15676,
  title  = {Dynamic Chunking for Diffusion Language Models},
  author = {Yichen Zhu and Xiaoming Shi and Peng Zhao and Weiyu Chen and Debing Zhang and James Kwok},
  journal= {arXiv preprint arXiv:2605.15676},
  year   = {2026}
}