arXiv 2406.111092024-11-16cs.CLcs.AIcs.LG

调查用于仇恨言论检测的大型语言模型中的标注者偏见

Investigating Annotator Bias in Large Language Models for Hate Speech Detection

University of North Alabama（北阿拉巴马大学）；Murray State University（默里州立大学）；Wichita State University（威奇托州立大学）；California State Polytechnic University Pomona（加州州立理工大学波莫纳分校）；Auburn University at Montgomery（奥本大学蒙哥马利分校）；Auburn University（奥本大学）；Gustavus Adolphus College（古斯塔夫斯·阿道弗斯学院）；Stanford University（斯坦福大学）；Amazon GenAI（亚马逊生成式人工智能部门）

Amit Das, Zheng Zhang, Najib Hasan, Souvika Sarkar, Fatemeh Jamshidi, Tathagata Bhattacharya, Mostafa Rahgouy, Nilanjana Raychawdhary, Dongji Feng, Vinija Jain,… 展开作者

Amit Das, Zheng Zhang, Najib Hasan, Souvika Sarkar, Fatemeh Jamshidi, Tathagata Bhattacharya, Mostafa Rahgouy, Nilanjana Raychawdhary, Dongji Feng, Vinija Jain, Aman Chadha, Mary Sandage, Lauramarie Pope, Gerry Dozier, Cheryl Seals

备注：: Accepted at NeurIPS Safe Generative AI Workshop, 2024

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文研究了GPT-3.5、GPT-4o、Llama-3.1和Gemma-2四个LLMs在仇恨言论数据标注中于性别、种族、宗教和残疾四类偏见，引入HateBiasNet数据集并在ETHOS上验证，为利用LLMs进行数据标注提供指导。

摘要：

数据标注，即为原始数据分配描述性标签的实践，在优化机器学习模型的性能中起着关键作用。然而，这是一个资源密集型的过程，容易受到标注者引入的偏见的影响。复杂的大型语言模型（LLMs）的出现为现代化和简化这一复杂流程提供了独特的机会。虽然现有研究广泛评估了LLMs作为标注者的有效性，但本文深入探讨了LLMs在标注仇恨言论数据时存在的偏见。我们的研究有助于理解四个关键类别中的偏见：性别、种族、宗教和残疾，并使用了四个LLMs：GPT-3.5、GPT-4o、Llama-3.1和Gemma-2。我们针对这些类别中高度脆弱的群体，分析了标注者偏见。此外，我们通过仔细审查标注数据，对促成这些偏见的潜在因素进行了全面检查。我们引入了定制的仇恨言论检测数据集HateBiasNet来进行这项研究。此外，我们还在ETHOS（Mollas等人，2022）数据集上进行了相同的实验以进行比较分析。本文作为一项关键资源，指导研究人员和从业者利用LLMs的潜力进行数据标注，从而促进这一关键领域的进步。

英文摘要：

Data annotation, the practice of assigning descriptive labels to raw data, is pivotal in optimizing the performance of machine learning models. However, it is a resource-intensive process susceptible to biases introduced by annotators. The emergence of sophisticated Large Language Models (LLMs) presents a unique opportunity to modernize and streamline this complex procedure. While existing research extensively evaluates the efficacy of LLMs, as annotators, this paper delves into the biases present in LLMs when annotating hate speech data. Our research contributes to understanding biases in four key categories: gender, race, religion, and disability with four LLMs: GPT-3.5, GPT-4o, Llama-3.1 and Gemma-2. Specifically targeting highly vulnerable groups within these categories, we analyze annotator biases. Furthermore, we conduct a comprehensive examination of potential factors contributing to these biases by scrutinizing the annotated data. We introduce our custom hate speech detection dataset, HateBiasNet, to conduct this research. Additionally, we perform the same experiments on the ETHOS (Mollas et al. 2022) dataset also for comparative analysis. This paper serves as a crucial resource, guiding researchers and practitioners in harnessing the potential of LLMs for data annotation, thereby fostering advancements in this critical field.

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。