REX-RAG: Reasoning Exploration with Policy Correction in Retrieval-Augmented Generation

Open in new window