Just on Time: Token-Level Early Stopping for Diffusion Language Models

arXiv:2602.11133v2 Announce Type: replace-cross Abstract: Diffusion language models generate text through iterative refinement, a process that is often computationally inefficient because many tokens reach stability long before the final denoising step. We introduce a training-free, token-level early stopping approach that identifies convergence independently at each position. Our method leverages lightweight signals derived from the model's predictions and local context to dynamically determine when individual tokens can be finalized. This yields adaptive per-token freezing without task-specific fine-tuning, substantially reducing the total number of diffusion steps required. Across diverse benchmarks, spanning mathematical reasoning, general question answering, and scientific understanding, our approach achieves substantial efficiency gains while preserving generation quality.

Related

• Stopping Computation for Converged Tokens in Masked Diffusion-LM Decoding
• Efficient Diffusion LLMs via Temporal-Spatial Parallel Decoding and Confidence Extrapolation
• DiffAdapt: Difficulty-Adaptive Reasoning for Token-Efficient LLM Inference

Source: arXiv cs.CL | 2026-08-04