arXivJunyi Wu, Tianchen Zhao, Shaoqiu Zhang, Linfeng Zhang, Guohao Dai, Yu WangMon, May 18, 2026, 3:09 AM PDT
score 17.0
Speeding up parallel text generation by compressing redundant mask tokens
Original: Elastic-dLLM: Position Preserving Context Compression and Augmentation of Diffusion LLMs
Source: arxiv.org ↗
Writing ELI5 summary…