多轮交互中大型语言模型的置信度估计

摘要

尽管置信度估计是缓解大语言模型幻觉现象的重要方向，但现有研究主要集中于单轮对话场景。在多轮对话中，随着上下文累积和歧义逐步消解，模型置信度的动态变化机制仍属研究空白。可靠的多轮置信度估计对自主智能体和人在回路的系统等下游应用至关重要。本研究首次对多轮交互中的置信度估计进行系统性探索，建立了基于双重核心诉求的正式评估框架：单轮校准性以及信息递增时置信度的单调性。为此我们提出了创新指标（包括长度归一化的预期校准误差InfoECE）和受控评估数据集生成的"提示者-猜测者"新范式。实验表明，主流置信度技术在多轮对话中难以保持校准性和单调性。我们提出的基于逻辑概率的探测方法P(Sufficient)取得了相对更优的性能，但该任务远未彻底解决。本研究为开发更可靠、可信的对话智能体奠定了方法论基础。

English

While confidence estimation is a promising direction for mitigating hallucinations in Large Language Models (LLMs), current research dominantly focuses on single-turn settings. The dynamics of model confidence in multi-turn conversations, where context accumulates and ambiguity is progressively resolved, remain largely unexplored. Reliable confidence estimation in multi-turn settings is critical for many downstream applications, such as autonomous agents and human-in-the-loop systems. This work presents the first systematic study of confidence estimation in multi-turn interactions, establishing a formal evaluation framework grounded in two key desiderata: per-turn calibration and monotonicity of confidence as more information becomes available. To facilitate this, we introduce novel metrics, including a length-normalized Expected Calibration Error (InfoECE), and a new "Hinter-Guesser" paradigm for generating controlled evaluation datasets. Our experiments reveal that widely-used confidence techniques struggle with calibration and monotonicity in multi-turn dialogues. We propose P(Sufficient), a logit-based probe that achieves comparatively better performance, although the task remains far from solved. Our work provides a foundational methodology for developing more reliable and trustworthy conversational agents.

多轮交互中大型语言模型的置信度估计

Confidence Estimation for LLMs in Multi-turn Interactions

摘要

Support