TY  - RPRT
TI  - GRACE: Reinforcement Learning for Grounded Response and Abstention under Contextual Evidence
AU  - Yibo Zhao
AU  - Jiapeng Zhu
AU  - Zichen Ding
AU  - Xiang Li
PY  - 2026
UR  - https://arxiv.org/abs/2601.04525
ID  - 2601.04525
ER  -