While Large Language Models (LLMs) show remarkable capabilities, their unreliability remains a critical barrier to deployment in high-stakes domains. This survey charts a functional evolution in addressing this challenge: the evolution of uncertainty from a passive diagnostic metric to an active control signal guiding real-time model behavior. We demonstrate how uncertainty is leveraged as an active control signal across three frontiers: in \textbf{advanced reasoning} to optimize computation and trigger self-correction; in \textbf{autonomous agents} to govern metacognitive decisions about tool use and information seeking; and in \textbf{reinforcement learning} to mitigate reward hacking and enable self-improvement via intrinsic rewards. By grounding these advancements in emerging theoretical frameworks like Bayesian methods and Conformal Prediction, we provide a unified perspective on this transformative trend. This survey provides a comprehensive overview, critical analysis, and practical design patterns, arguing that mastering the new trend of uncertainty is essential for building the next generation of scalable, reliable, and trustworthy AI.

翻译：虽然大语言模型（LLMs）展现出卓越的能力，但其不可靠性仍然是部署于高风险领域的关键障碍。本综述描绘了应对这一挑战的功能性演变：不确定性从一种被动的诊断度量指标进化为一种指导实时模型行为的主动控制信号。我们展示了不确定性如何作为主动控制信号在三个前沿方向被利用：在**高级推理**中，用于优化计算并触发自我修正；在**自主智能体**中，用于管理关于工具使用和信息搜寻的元认知决策；以及在**强化学习**中，用于减轻奖励篡改并通过内在奖励实现自我改进。通过将这些进展扎根于新兴的理论框架，如贝叶斯方法和共形预测，我们为这一变革性趋势提供了统一的视角。本综述提供了全面的概述、批判性的分析以及实用的设计模式，主张掌握不确定性的新趋势对于构建下一代可扩展、可靠且值得信赖的人工智能至关重要。