大型语言模型真正意义上的扎根（grounding）能力究竟如何？
How Well Do Large Language Models Truly Ground?
AI总结：
本文提出更严格的LLM扎根定义，并构建新数据集与指标，在25个模型上实验分析影响扎根能力的因素，以提升LLM应用的可靠性与可控性。
摘要：
为了减少大型语言模型（LLMs）中的幻觉和缺乏可控性等问题，一种常见方法是基于作为输入提供的外部上下文来生成回答，这类模型被称为知识增强模型。然而，以往研究往往将“扎根”狭义地定义为仅仅给出正确答案，这并不能保证整个回答的可靠性。为克服这一问题，我们提出了一个更严格的扎根定义：一个模型只有在（1）充分利用所提供上下文中的必要知识，并且（2）始终不超出该知识范围时，才算是真正扎根。我们引入了一个新的数据集和一个扎根评估指标，用于在该定义下评估模型能力。我们在25个不同规模和不同训练方法的LLMs上进行了实验，并提供了关于影响扎根性能因素的分析。我们的发现有助于更好地理解如何提升扎根能力，并为朝着更可靠、更可控的LLM应用方向发展指出了一个改进方向。
英文摘要：
To reduce issues like hallucinations and lack of control in Large Language Models (LLMs), a common method is to generate responses by grounding on external contexts given as input, known as knowledge-augmented models. However, previous research often narrowly defines "grounding" as just having the correct answer, which does not ensure the reliability of the entire response. To overcome this, we propose a stricter definition of grounding: a model is truly grounded if it (1) fully utilizes the necessary knowledge from the provided context, and (2) stays within the limits of that knowledge. We introduce a new dataset and a grounding metric to evaluate model capability under the definition. We perform experiments across 25 LLMs of different sizes and training methods and provide insights into factors that influence grounding performance. Our findings contribute to a better understanding of how to improve grounding capabilities and suggest an area of improvement toward more reliable and controllable LLM applications.
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。
