MARS:生成式大语言模型不确定性估计的意义感知响应评分
MARS: Meaning-Aware Response Scoring for Uncertainty Estimation in Generative LLMs
AI总结:
针对生成式LLM输出的不确定性估计问题,提出意义感知响应评分(MARS)函数替代长度归一化评分,在多数据集和多LLM上实验证明可显著提升UE性能,并在医学QA数据集验证有效性。
摘要:
生成式大型语言模型(LLMs)因其在各类任务中的出色表现而被广泛应用。然而,它们生成不准确或误导性输出的倾向带来了潜在风险,尤其在高风险环境中。因此,评估生成式LLM输出的正确性是提升可靠性的重要任务。生成式LLM的不确定性估计(UE)是一个不断发展的领域,当前基于概率的最先进(SOTA)方法通常采用长度归一化评分。在本研究中,我们提出意义感知响应评分(MARS)作为UE方法中长度归一化评分的替代方案。MARS是一种新颖的评分函数,在问题语境中考虑生成序列中每个标记的语义贡献。我们证明,将MARS集成到UE方法中可实现UE性能的普遍且显著提升。我们使用三个不同的闭卷问答数据集在五个流行的预训练LLM上进行了实验。最后,我们在医学问答数据集上验证了MARS的有效性。代码可在https://github.com/Ybakman/LLM_Uncertainity获取。
英文摘要:
Generative Large Language Models (LLMs) are widely utilized for their excellence in various tasks. However, their tendency to produce inaccurate or misleading outputs poses a potential risk, particularly in high-stakes environments. Therefore, estimating the correctness of generative LLM outputs is an important task for enhanced reliability. Uncertainty Estimation (UE) in generative LLMs is an evolving domain, where SOTA probability-based methods commonly employ length-normalized scoring. In this work, we propose Meaning-Aware Response Scoring (MARS) as an alternative to length-normalized scoring for UE methods. MARS is a novel scoring function that considers the semantic contribution of each token in the generated sequence in the context of the question. We demonstrate that integrating MARS into UE methods results in a universal and significant improvement in UE performance. We conduct experiments using three distinct closed-book question-answering datasets across five popular pre-trained LLMs. Lastly, we validate the efficacy of MARS on a Medical QA dataset. Code can be found https://github.com/Ybakman/LLM_Uncertainity.
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。