SIRAG: Towards Stable and Interpretable RAG with A Process-Supervised Multi-Agent Framework

Open in new window