New paper proposes FEROM method for masked diffusion language models
An arXiv paper introduces FEROM, a marginal policy optimization approach for masked diffusion language models that addresses how samplers select reveal positions when iteratively unmasking text.
Read the full story at arXiv cs.CL ↗
Timeline · 1 report
- 2026-11-10 05:00 · arXiv cs.CL
FEROM: Frontier Endogenous Reveal-Order Marginal Policy Optimization for Masked Diffusion LMs