CoT-UQ：利用思维链提升大语言模型响应级不确定性量化
CoT-UQ: Improving Response-wise Uncertainty Quantification in LLMs with Chain-of-Thought
AI总结：
本文提出CoT-UQ框架，将思维链推理中的关键词重要性信息聚合用于响应级不确定性量化，在Llama 8B至13B模型上较现有UQ方法平均提升5.9% AUROC。
摘要：
大语言模型（LLMs）在许多任务中表现出色，但难以准确量化其生成响应中的不确定性。这一局限使得检测错误信息和确保可靠决策变得困难。现有的大语言模型不确定性量化（UQ）方法主要是提示级而非响应级的，通常需要多个响应样本，从而带来高昂的计算成本。此外，已有研究表明大语言模型往往过度自信，尤其是在使用推理步骤得出答案时。在本工作中，我们提出CoT-UQ，一个响应级UQ框架，将大语言模型通过思维链（Chain-of-Thought，CoT）展现的内在推理能力整合到UQ过程中。CoT-UQ在推理过程中通过从每个推理步骤中提取关键词并评估其对最终答案的重要性来捕获关键信息。随后，这些关键推理信息被聚合以产生最终的不确定性估计。我们基于Llama系列模型进行了广泛实验，模型规模从8B到13B不等，涵盖逻辑推理和数学推理任务。实验结果表明，CoT-UQ显著优于现有UQ方法，与当前UQ方法相比，AUROC平均提升5.9%。代码可在以下地址获取：https://github.com/ZBox1005/CoT-UQ。
英文摘要：
Large language models (LLMs) excel in many tasks but struggle to accurately quantify uncertainty in their generated responses. This limitation makes it challenging to detect misinformation and ensure reliable decision-making. Existing uncertainty quantification (UQ) methods for LLMs are primarily prompt-wise rather than response-wise, often requiring multiple response samples, which incurs high computational costs. Moreover, LLMs have been shown to be overconfident, particularly when using reasoning steps to derive their answers. In this work, we propose CoT-UQ, a response-wise UQ framework that integrates LLMs' inherent reasoning capabilities through Chain-of-Thought (CoT) into the UQ process. CoT-UQ captures critical information during inference by extracting keywords from each reasoning step and assessing their importance to the final answer. This key reasoning information is then aggregated to produce a final uncertainty estimate. We conduct extensive experiments based on Llama Family with model sizes varying from 8B to 13B across logical and mathematical reasoning tasks. Experimental results demonstrate that CoT-UQ significantly outperforms existing UQ methods, achieving an average improvement of 5.9% AUROC compared to current UQ methods. The code is available at: https://github.com/ZBox1005/CoT-UQ.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
