AI.info

Diyi Hu

Explore Diyi Hu on AI.info.

• ProbePO: Sparse Probes, Redistributed Credit for Critic-Free LLM Reinforcement Learning