Skip to content
NeurIPS 2026 planner

Poster · Sydney

VESPO: Variational Sequence-level Soft Policy Optimization for Off-Policy LLM Training

When
Tue 8 Dec 2026, 17:00–20:00 AEDT
Where
Hall 1-4, ICC Sydney
Exhibition Centre, Level 1
Session
Sydney Poster Session 2
Authors
Guobin Shen, Chenxiao Zhao, Xiang Cheng, Lei Huang, XingYu Li
Organisations
ByteDance, Xiaohongshu, Harbin Institute of Technology
QRS classification
Reinforcement learning, Language models, Probabilistic methods

Open in scheduleOfficial page