TY  - RPRT
TI  - STEVE: Stabilizing Textual Gradient-Based Prompt Optimization via Error-Driven Refinement and Regularized Verification
AU  - Yifan Xu
AU  - Yixuan Li
AU  - Xinzhuo Li
AU  - Yixin Gu
AU  - Yifan Shen
AU  - Lijun Yu
AU  - Haohan Wang
PY  - 2026
UR  - https://arxiv.org/abs/2609.23716
ID  - 2609.23716
ER  -