5 citations · 7 across the 3 of their papers we have counts for
Showing cs.LGShow all
2 papers · 1 filter
cs.LG2026
Outcome-Routed Distillation and Reinforcement Learning for Diffusion Language Models
Wentao Lu, Tianyu Zhu, Jesse Clark
On-policy distillation provides dense supervision on the student's own samples, but its targets still come from the teacher. Verifier reinforcement learning (RL) instead uses an ou…
cs.LG2026
Context-Tower Conversion Preserves Generation While Freezing Retains Knowledge: Low-Budget AR-to-Diffusion Conversion of MoE LLMs
Wentao Lu, Jesse Clark, Tianyu Zhu
Converting a pretrained autoregressive (AR) model to a diffusion language model (dLLM) enables parallel generation without pretraining a new model. Published conversion methods dif…