SAID: Accelerating Diffusion-Based Language Models via Scaffold-Aware Iterative Decoding

Add code
Jun 03, 2026

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: