LLM-as-a-Judge中的自我偏好偏差
Self-Preference Bias in LLM-as-a-Judge
AI总结：
本文提出一种定量指标测量LLM-as-a-judge中的自我偏好偏差，发现GPT-4存在显著偏差，并揭示其本质源于对低困惑度（更熟悉）文本的偏好。
摘要：
利用大型语言模型（LLM）进行的自动化评估，通常称为LLM评估器或LLM-as-a-judge，已被广泛用于衡量对话系统的性能。然而，LLM中的自我偏好偏差带来了显著风险，包括推广LLM固有的特定风格或策略。尽管该问题具有重要性，但目前缺乏定量测量自我偏好偏差的既定方法，且其根本原因尚不明确。在本文中，我们引入了一种新颖的定量指标来测量自我偏好偏差。我们的实验结果表明，GPT-4表现出显著的自我偏好偏差。为了探究原因，我们假设LLM可能偏好那些对它们更熟悉的输出，这通过较低的困惑度来体现。我们分析了LLM评估与输出困惑度之间的关系。我们的发现揭示，无论输出是否为自我生成，LLM对较低困惑度输出的评估显著高于人类评估者。这表明偏差的本质在于困惑度，而自我偏好偏差的存在是因为LLM偏好对它们更熟悉的文本。
英文摘要：
Automated evaluation leveraging large language models (LLMs), commonly referred to as LLM evaluators or LLM-as-a-judge, has been widely used in measuring the performance of dialogue systems. However, the self-preference bias in LLMs has posed significant risks, including promoting specific styles or policies intrinsic to the LLMs. Despite the importance of this issue, there is a lack of established methods to measure the self-preference bias quantitatively, and its underlying causes are poorly understood. In this paper, we introduce a novel quantitative metric to measure the self-preference bias. Our experimental results demonstrate that GPT-4 exhibits a significant degree of self-preference bias. To explore the causes, we hypothesize that LLMs may favor outputs that are more familiar to them, as indicated by lower perplexity. We analyze the relationship between LLM evaluations and the perplexities of outputs. Our findings reveal that LLMs assign significantly higher evaluations to outputs with lower perplexity than human evaluators, regardless of whether the outputs were self-generated. This suggests that the essence of the bias lies in perplexity and that the self-preference bias exists because LLMs prefer texts more familiar to them.
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。
