让大模型基于自身不确定性判断何时回答、澄清或放弃。
Clarify, Abstain or Answer? Strategising in Conversation with Belief-Augmented Generation
- 用模型自生的多种回应构建信念状态，指导对话策略选择。
- 在六种模型上提升问答准确率，决策更符合真实不确定性。
- 适合需要可靠推理与风险控制的对话系统开发者。
大型语言模型（LLMs）生成文本时隐含一种概率分布，可视为对不确定性的表示：采样K个回复即构成模型认为合理的信念状态。现有方法仅用于特定任务如解码或选择性预测，且常需人工干预，无法直接控制生成过程。本文提出信念增强生成（Belief-Augmented Generation, BAG）：通过提示词将模型锚定在其自身的信念状态，使其基于这K个样本进行推理，决定对话策略——回答、澄清或放弃。在多轮模糊问答场景中，我们发现模型默认很少澄清或放弃，忽视输入或事实的不确定性。BAG在六种模型上均提升了问答准确率，并使策略决策更贴近信念状态，但区分‘何时澄清’与‘何时放弃’仍具挑战。
原文摘要 · Abstract (English)
Large language models (LLMs) define a distribution over text, which can be viewed as a probabilistic representation of uncertainty: sampling K responses yields a belief state - responses a model deems plausible. Existing work exploits this representation for narrow tasks like either decoding or selective prediction, and often requires manual interventions, not controlling generation directly. We propose Belief-Augmented Generation (BAG): grounding LLMs in their own belief state via the prompt and letting them reason over these K samples to decide on a conversational strategy: answer, clarify, or abstain. In a multi-turn ambiguous QA setting, we find that LLMs by default rarely clarify or abstain, ignoring uncertainty about the input or facts. BAG improves QA accuracy across six models and yields strategy decisions more faithful to the belief state than prompt-only baselines. Disentangling when to clarify from when to abstain, however, remains challenging.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。