ORL is introduced, a framework for on-policy residual learning in DLMs that learns from model-induced residual trajectories on the fly and introduces an entropy-contraction objective that sharpens predictions as the model revisits its own predictions.
Yi-Fan Yu, Yu-Qing Jian, Ji-Sen Li et al.· 0 citations
Speculative decoding is critical for accelerating LLM inference. However, the speedup is fragile: drafters are typically trained against a narrow distribution for a single target model, and their acceptance rate collapses under workload shifts. This is a striking inversion of modern LLM development, where target models...
Fengxiang Bie, Yu-Qing Jian, Yi-Fan Yu et al.· 0 citations
Reinforcement learning with verifiable rewards (RLVR) is rapidly advancing the reasoning capabilities of language models, yet the optimization layer that converts reward feedback into weight-space updates remains poorly understood. Building on our prior analysis (Zhu et al., 2025), we study this missing layer through t...