04:00
2026-08-05
arxiv.org
large-language-models
Speculative Correction: Draft-then-Refine Decoding for Diffusion Language Models
A new arXiv preprint (arXiv:2608.02625v1) introduces draft-then-refine decoding for diffusion language models, showing that LLaDA2.1-Flash improves GSM8K-384 accuracy from 0.848 to 0.899 and MBPP-384 …