PACT: Certifying What Is Written in Parallel Decoding of Diffusion Language Models
Parallel decoding speeds up masked diffusion language models by writing several tokens per forward pass. Published commit rules choose these tokens by testing the predicted distributions. Under sampling the written tokens are drawn, and these tests do not see them. We find that answer quality follows the reliability of...
Yan-Chao Li, Yu-Su Zhang, Jia-Qing Xie et al.
· 0 citations