Dual-Channel Robust Group-Relative Policy Optimization via Advantage and Sequence-Weight Estimation
RoVR-GSPO is introduced, a dual-channel robust optimizer that addresses group-relative policy optimization failure modes separately and shows consistent improvements over GSPO, while controlled perturbation studies demonstrate stronger robustness to reward contamination and token-ratio anomalies.
Zhong-Yi Li, Wan Tian, Xiang Xu et al.
· 0 citations