TY  - RPRT
TI  - Breaking Contextual Inertia: Reinforcement Learning with Single-Turn Anchors for Stable Multi-Turn Interaction
AU  - Xingwu Chen
AU  - Zhanqiu Zhang
AU  - Yiwen Guo
AU  - Difan Zou
PY  - 2026
UR  - https://arxiv.org/abs/2603.04783
ID  - 2603.04783
ER  -