arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2402.01740cs.CLcs.AI
收藏

减少大型语言模型中的选择偏差

Reducing Selection Bias in Large Language Models

J. E. Eicher, R. F. Irgolič

AI总结:

本文通过操控温度、列表长度、对象类型等因素,实证研究了LLM在列表选择任务中的偏差结构,发现首因效应和护栏提示会加剧偏差,而将护栏与采样步骤分离可消除偏差,并提出LLM以偏差补偿认知负荷的理论。

摘要:

像 gpt-3.5-turbo-0613 和 claude-instant-1.2 这样的大型语言模型(LLMs)在解释和执行语义任务中至关重要。遗憾的是,这些模型固有的偏差会对其性能产生不利影响,尤其受影响的是从列表中选择对象这一操作;这是数字导航和决策中的一项基本操作。本研究批判性地考察了这些偏差,并量化了它们对代表性列表选择任务的影响。为了探究这些偏差,我们通过操控温度、列表长度、对象身份、对象类型、提示复杂度和模型进行了实验。我们分离并测量了这些偏差对选择行为的影响。我们的研究结果表明,偏差结构强烈依赖于模型,而对象类型调节了效应的大小。存在强烈的首因效应,导致列表中的前几个对象在输出中被不成比例地表示。使用护栏(guard rails)——一种确保响应结构的提示工程方法——在应用于选择任务时,会增加偏差并降低指令遵循度。当将护栏步骤与列表采样步骤分离,降低每个单独任务的复杂度时,该偏差会被消除。我们提供了 LLM 应用建议,并从理论上提出 LLM 经历了一种认知负荷形式,而这种认知负荷通过偏差得到了补偿。

英文摘要:

Large Language Models (LLMs) like gpt-3.5-turbo-0613 and claude-instant-1.2 are vital in interpreting and executing semantic tasks. Unfortunately, these models' inherent biases adversely affect their performance Particularly affected is object selection from lists; a fundamental operation in digital navigation and decision-making. This research critically examines these biases and quantifies the effects on a representative list selection task. To explore these biases, we experiment manipulating temperature, list length, object identity, object type, prompt complexity, and model. We isolated and measured the influence of the biases on selection behavior. Our findings show that bias structure is strongly dependent on the model, with object type modulating the magnitude of the effect. With a strong primacy effect, causing the first objects in a list to be disproportionately represented in outputs. The usage of guard rails, a prompt engineering method of ensuring a response structure, increases bias and decreases instruction adherence when to a selection task. The bias is ablated when the guard rail step is separated from the list sampling step, lowering the complexity of each individual task. We provide LLM applications and theoretically suggest that LLMs experience a form of cognitive load that is compensated for with bias.

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。

↑