大型语言模型中的默许偏差
Acquiescence Bias in Large Language Models
AI总结：
本研究调查大型语言模型中的默许偏差，发现与人类不同，LLMs倾向于回答“否”，无论其表示同意或不同意，且该现象跨模型、任务和语言存在。
摘要：
默许偏差，即人类在调查中倾向于同意陈述，而不论其实际信念如何，这一现象已被充分研究和记录。鉴于大型语言模型（LLMs）已被证明极易受到输入中相对较小变化的影响，并且是在人类生成的数据上进行训练的，因此合理推测它们可能表现出类似的倾向。我们提出了一项研究，调查LLMs在不同模型、任务和语言（英语、德语和波兰语）中是否存在默许偏差。我们的结果表明，与人类相反，LLMs表现出倾向于回答“否”的偏差，无论该回答表示同意还是不同意。
英文摘要：
Acquiescence bias, i.e. the tendency of humans to agree with statements in surveys, independent of their actual beliefs, is well researched and documented. Since Large Language Models (LLMs) have been shown to be very influenceable by relatively small changes in input and are trained on human-generated data, it is reasonable to assume that they could show a similar tendency. We present a study investigating the presence of acquiescence bias in LLMs across different models, tasks, and languages (English, German, and Polish). Our results indicate that, contrary to humans, LLMs display a bias towards answering no, regardless of whether it indicates agreement or disagreement.
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。
