TY  - RPRT
TI  - DeCoRL: Decoupling Reasoning Chains via Parallel Sub-Step Generation and Cascaded Reinforcement for Interpretable and Scalable RLHF
AU  - Ziyuan Gao
AU  - Di Liang
AU  - Xianjie Wu
AU  - Philippe Morel
AU  - Minlong Peng
PY  - 2025
UR  - https://arxiv.org/abs/2511.19097
ID  - 2511.19097
ER  -