InternReviewer and InternAdvocate for agentic scholarly RL

This framework introduces specialized agents for peer review and rebuttal using an agentic Reinforcement Learning paradigm. It utilizes a multi-dimensional reward system and real-time citation verification to minimize hallucinations and subjective model bias.