Mitigating Factual Hallucination in LRMs via Mixed-Mode Advantage Regularization

2 points | by root-parent 9 hours ago