AI.info

Minlong Peng

Explore Minlong Peng on AI.info.

• DeCoRL: Decoupling Reasoning Chains via Parallel Sub-Step Generation and Cascaded Reinforcement for Interpretable and Scalable RLHF