事实性增强对LLM是免费午餐吗？更好的事实性可能导致更差的上下文忠实度
Is Factuality Enhancement a Free Lunch For LLMs? Better Factuality Can Lead to Worse Context-Faithfulness
AI总结：
本文研究了LLM事实性增强方法对上下文忠实度的负面影响，通过实验和隐藏状态分析发现这些方法虽可能提升事实准确性，却会导致上下文忠实度显著下降，最大降幅达69.7%，并呼吁未来研究在增强事实性的同时减少对上下文忠实度的牺牲。
摘要：
作为现代文本理解和生成的首选工具，大语言模型（LLMs）被期望能够利用输入上下文准确输出答案。这要求LLMs同时具备上下文忠实度和事实准确性。人们已经做出了大量努力，通过事实性增强方法来缓解幻觉，从而使LLMs产生更好的输出。然而，这些方法也带来了阻碍上下文忠实度的风险，因为事实性增强可能导致LLMs对其参数化知识过于自信，从而忽视相关的输入上下文。在这项工作中，我们认为当前的事实性增强方法会显著削弱LLMs的上下文忠实度。我们首先重新审视当前的事实性增强方法，并评估它们在增强事实准确性方面的有效性。接着，我们评估它们在知识编辑任务上的表现，以评估对上下文忠实度的潜在影响。实验结果表明，虽然这些方法在事实准确性方面可能带来不一致的提升，但它们也会导致上下文忠实度更严重的下降，最大降幅达到惊人的69.7%。为了解释这些下降，我们分别分析了代表新知识和参数化知识的token的隐藏状态和logit分布，揭示了当前方法的局限性。我们的发现凸显了增强LLMs时固有的复杂权衡。因此，我们建议更多关于LLMs事实性增强的研究应努力减少对上下文忠实度的牺牲。
英文摘要：
As the modern tools of choice for text understanding and generation, large language models (LLMs) are expected to accurately output answers by leveraging the input context. This requires LLMs to possess both context-faithfulness and factual accuracy. Extensive efforts have been made to enable better outputs from LLMs by mitigating hallucinations through factuality enhancement methods. However, they also pose risks of hindering context-faithfulness, as factuality enhancement can lead LLMs to become overly confident in their parametric knowledge, causing them to overlook the relevant input context. In this work, we argue that current factuality enhancement methods can significantly undermine the context-faithfulness of LLMs. We first revisit the current factuality enhancement methods and evaluate their effectiveness in enhancing factual accuracy. Next, we evaluate their performance on knowledge editing tasks to assess the potential impact on context-faithfulness. The experimental results reveal that while these methods may yield inconsistent improvements in factual accuracy, they also cause a more severe decline in context-faithfulness, with the largest decrease reaching a striking 69.7\%. To explain these declines, we analyze the hidden states and logit distributions for the tokens representing new knowledge and parametric knowledge respectively, highlighting the limitations of current approaches. Our finding highlights the complex trade-offs inherent in enhancing LLMs. Therefore, we recommend that more research on LLMs' factuality enhancement make efforts to reduce the sacrifice of context-faithfulness.
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。
