Scaling Large Reasoning Models beyond Human Supervision: A Path toward Superintelligence Paper โข 2608.31075 โข Published 14 days ago โข 31