你能信任LLM的判断吗？LLM作为评判者的可靠性
Can You Trust LLM Judgments? Reliability of LLM-as-a-Judge
AI总结：
针对LLM随机性导致判断不可靠的问题，本文提出利用麦克唐纳ω评估LLM作为评判者可靠性的框架，通过单/多轮基准实验分析温度影响，证明多样本评估的必要性，为构建可信LLM系统提供关键支撑。
摘要：
大型语言模型（LLMs）已变得越来越强大且无处不在，但其随机性对输出可靠性构成挑战。尽管确定性设置可提高一致性，但无法保证可靠性，因为模型概率分布中的单个样本仍可能具有误导性。基于LLM作为评判者的概念，我们引入了一个利用麦克唐纳ω（McDonald's omega）严格评估LLM判断可靠性的新框架。我们在标准单轮和多轮基准上评估LLMs评判其他LLMs输出时的可靠性，同时研究温度对可靠性的影响。通过分析这些结果，我们证明了固定随机性的局限性以及考虑多个样本的重要性，这对下游应用具有重要意义。我们的发现强调了需要细致理解LLM可靠性，以及过度依赖单样本评估相关的潜在风险。这项工作为构建更可信、更可靠的基于LLM的系统和应用迈出了关键一步。
英文摘要：
Large Language Models (LLMs) have become increasingly powerful and ubiquitous, but their stochastic nature poses challenges to the reliability of their outputs. While deterministic settings can improve consistency, they do not guarantee reliability, as a single sample from the model's probability distribution can still be misleading. Building upon the concept of LLM-as-a-judge, we introduce a novel framework for rigorously evaluating the reliability of LLM judgments, leveraging McDonald's omega. We evaluate the reliability of LLMs when judging the outputs of other LLMs on standard single-turn and multi-turn benchmarks, simultaneously investigating the impact of temperature on reliability. By analyzing these results, we demonstrate the limitations of fixed randomness and the importance of considering multiple samples, which we show has significant implications for downstream applications. Our findings highlight the need for a nuanced understanding of LLM reliability and the potential risks associated with over-reliance on single-shot evaluations. This work provides a crucial step towards building more trustworthy and reliable LLM-based systems and 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。
