利用可解释人工智能进行 LLM 文本归属：区分人类撰写文本与多个 LLM 生成文本
Leveraging Explainable AI for LLM Text Attribution: Differentiating Human-Written and Multiple LLMs-Generated Text
AI总结：
本文结合 RF、RNN 等机器学习模型与 XAI，识别人类文本及五种 LLM 生成文本，二分类和多分类表现优异，并以 98.5% 准确率显著优于 GPTZero。
摘要：
生成式人工智能大语言模型（LLM）的发展，为识别由生成式 AI 或人类产生的内容敲响了警钟。在一种情形中，当学生过度依赖此类工具、以至可能影响其写作或编码技能发展时，问题便会出现。其他剽窃问题也同样存在。本研究旨在支持检测和识别使用 LLM 工具生成的文本内容。我们假设 LLM 生成文本可被机器学习（ML）检测，并研究能够识别和区分由多个 LLM 工具生成文本的 ML 模型。我们利用多种 ML 和深度学习（DL）算法，例如 Random Forest（RF，随机森林）和 Recurrent Neural Networks（RNN，循环神经网络），并使用 Explainable Artificial Intelligence（XAI，可解释人工智能）来理解归属中的重要特征。我们的方法分为：1）二分类，用于区分人类撰写文本与 AI 文本；2）多分类，用于区分人类撰写文本与由五种不同 LLM 工具（ChatGPT、LLaMA、Google Bard、Claude 和 Perplexity）生成的文本。结果显示，多分类和二分类均具有较高准确率。我们的模型以 98.5% 对 78.3% 的准确率优于 GPTZero。值得注意的是，GPTZero 无法识别约 4.2% 的观测样本，而我们的模型能够识别完整的测试数据集。XAI 结果表明，理解不同类别之间的特征重要性能够形成详细的作者/来源画像。此外，这还有助于归属，并通过突出独特的风格和结构元素来支持剽窃检测，从而确保稳健的内容原创性验证。
英文摘要：
The development of Generative AI Large Language Models (LLMs) raised the alarm regarding identifying content produced through generative AI or humans. In one case, issues arise when students heavily rely on such tools in a manner that can affect the development of their writing or coding skills. Other issues of plagiarism also apply. This study aims to support efforts to detect and identify textual content generated using LLM tools. We hypothesize that LLMs-generated text is detectable by machine learning (ML), and investigate ML models that can recognize and differentiate texts generated by multiple LLMs tools. We leverage several ML and Deep Learning (DL) algorithms such as Random Forest (RF), and Recurrent Neural Networks (RNN), and utilized Explainable Artificial Intelligence (XAI) to understand the important features in attribution. Our method is divided into 1) binary classification to differentiate between human-written and AI-text, and 2) multi classification, to differentiate between human-written text and the text generated by the five different LLM tools (ChatGPT, LLaMA, Google Bard, Claude, and Perplexity). Results show high accuracy in the multi and binary classification. Our model outperformed GPTZero with 98.5\% accuracy to 78.3\%. Notably, GPTZero was unable to recognize about 4.2\% of the observations, but our model was able to recognize the complete test dataset. XAI results showed that understanding feature importance across different classes enables detailed author/source profiles. Further, aiding in attribution and supporting plagiarism detection by highlighting unique stylistic and structural elements ensuring robust content originality verification.
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。
