TY  - RPRT
TI  - Toward Plasticity-Preserving KL Regularization for Capability Retention in LLM Reinforcement Learning
AU  - Li Wang
AU  - Xiaodong Lu
AU  - Xiaohan Wang
AU  - Jiajun Chai
AU  - Wei Lin
AU  - Tianhao Peng
AU  - Guojun Yin
PY  - 2026
UR  - https://arxiv.org/abs/2608.01743
ID  - 2608.01743
ER  -