STDec: Spatio-Temporal Stability Guided Decoding for dLLMs

원문 초록

Diffusion Large Language Models (dLLMs) have achieved rapid progress, viewed as a promising alternative to the autoregressive paradigm. However, most dLLM decoders still adopt a global confidence threshold, and do not explicitly model local context from neighboring decoded states or temporal consistency of predicted token IDs across steps. To address this issue, we propose a spatio-temporal stability guided decoding approach, named STDec. We observe strong spatio-temporal stability in dLLM decoding: newly decoded tokens tend to lie near decoded neighbors, and their predicted IDs often remain consistent across several denoising steps. Inspired by this stability, our STDec introduces spatial-aware decoding and temporal-aware decoding. The spatial-aware decoding dynamically generates the token-adaptive thresholds by aggregating the decoded states of nearby tokens. The temporal-aware decoding relaxes the decoding thresholds for tokens whose predicted token IDs remain consistent over denoising steps. Our STDec is training-free and remains compatible with cache-based acceleration methods. Across textual reasoning and multimodal understanding benchmarks, STDec achieves improved inference throughput while maintaining comparable generation score. Notably, on MBPP with LLaDA, STDec achieves up to 14.17$\times$ speedup with a comparable score. The homepage is available at https://yzchen02.github.io/STDec.