前沿大语言模型中的记忆与版权侵权探究：2023年《纽约时报》诉OpenAI案研究
Exploring Memorization and Copyright Violation in Frontier LLMs: A Study of the New York Times v. OpenAI 2023 Lawsuit
AI总结：
本研究围绕2023年《纽约时报》诉OpenAI案，探究前沿LLM的逐字记忆倾向，发现超千亿参数模型记忆能力显著提升，OpenAI模型较难被诱导出记忆内容，为模型训练优化和AI版权法律问题提供参考。
摘要：
2023年12月《纽约时报》对OpenAI提起诉讼，使得前沿大语言模型（LLM）的版权侵权问题近期备受关注。《纽约时报》主张，GPT-4存在版权侵权行为：一是复制其文章用于LLM训练，二是记忆输入内容并在输出中公开展示这些内容。本研究旨在衡量OpenAI的LLM相较于其他LLM，在输出中呈现逐字记忆内容的倾向，研究重点聚焦于新闻文章。我们发现，GPT和Claude模型均采用了拒绝训练和输出过滤机制，以阻止记忆文章的逐字输出。我们使用基础提示模板绕过了拒绝训练，结果显示，当前OpenAI的模型相比Meta、Mistral和Anthropic的模型，更难被诱导出记忆内容。研究还发现，随着模型规模增大，尤其是参数量超过1000亿后，模型的记忆能力会显著提升。这些发现对模型训练具有实践意义：必须更加重视防止超大型模型出现逐字记忆问题。同时，这些发现也具有法律意义：通过评估OpenAI LLM的相对记忆能力，我们探讨了《纽约时报》版权侵权主张的合理性与OpenAI法律抗辩的力度，同时凸显了生成式AI与法律、政策交叉领域的相关问题。
英文摘要：
Copyright infringement in frontier LLMs has received much attention recently due to the New York Times v. OpenAI lawsuit, filed in December 2023. The New York Times claims that GPT-4 has infringed its copyrights by reproducing articles for use in LLM training and by memorizing the inputs, thereby publicly displaying them in LLM outputs. Our work aims to measure the propensity of OpenAI's LLMs to exhibit verbatim memorization in its outputs relative to other LLMs, specifically focusing on news articles. We discover that both GPT and Claude models use refusal training and output filters to prevent verbatim output of the memorized articles. We apply a basic prompt template to bypass the refusal training and show that OpenAI models are currently less prone to memorization elicitation than models from Meta, Mistral, and Anthropic. We find that as models increase in size, especially beyond 100 billion parameters, they demonstrate significantly greater capacity for memorization. Our findings have practical implications for training: more attention must be placed on preventing verbatim memorization in very large models. Our findings also have legal significance: in assessing the relative memorization capacity of OpenAI's LLMs, we probe the strength of The New York Times's copyright infringement claims and OpenAI's legal defenses, while underscoring issues at the intersection of generative AI, law, and policy.
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。
