PPL
Shuo Yang
Papers (6)
LATTE: Forecasting Peer Anchored Preference Trajectories for Personalized LLM GenerationBeyond the Target: From Imitation to Collaboration in Speculative DecodingClipping Bottleneck: Stabilizing RLVR via Stochastic Recovery of Near-Boundary SignalsOne-Way Policy Optimization for Self-Evolving LLMsReasoning Portability: Guiding Continual Learning for MLLMs in the RLVR EraActive Tabular Augmentation via Policy-Guided Diffusion Inpainting