T^star: Progressive Block Scaling for Masked Diffusion Language Models Through Trajectory Aware Reinforcement Learning

arXiv:2601.11214v5 Announce Type: replace Abstract: We present T^star, a simple TraceRL-based training curriculum for progressive block-size scaling in masked diffusion language models (MDMs). Starting from an AR-initialized small-block MDM, T^star transitions smoothly to larger blocks, enabling higher-parallelism decoding with minimal performance degradation on math reasoning benchmarks. Moreover, further analysis suggests that T^star may actually converge to an alternative decoding schedule that achieves comparable performance.

Source: arXiv cs.CL | 2026-06-04