Elephants Never Forget: Memorization and Learning of Tabular Data in Large Language Models
Sebastian Bordt University of Tübingen, Tübingen AI Center \ANDHarsha Nori、Vanessa Rodrigues、Besmira Nushi、Rich Caruana Microsoft Research
Abstract
While many have shown how Large Language Models (LLMs) can be applied to a diverse set of tasks, the critical issues of data contamination and memorization are often glossed over. In this work, we address this concern for tabular data. Specifically, we introduce a variety of different techniques to assess whether a language model has seen a tabular dataset during training. This investigation reveals that LLMs have memorized many popular tabular datasets verbatim. We then compare the few-shot learning performance of LLMs on datasets that were seen during training to the performance on datasets released after training. We find that LLMs perform better on datasets seen during training, indicating that memorization leads to overfitting. At the same time, LLMs show non-trivial performance on novel datasets and are surprisingly robust to data transformations. We then investigate the in-context statistical learning abilities of LLMs. While LLMs are significantly better than random at solving statistical classification problems, the sample efficiency of few-shot learning lags behind traditional statistical learning algorithms, especially as the dimension of the problem increases. This sugg
中文速览
大型语言模型（LLM）在表格数据上的"优异表现"到底有多少是真本事、又有多少是"背过答案"？研究者针对这一问题设计了四种检测方法，系统测试 GPT-3.5、GPT-4 等主流模型是否逐字记住了常用表格数据集，结果发现这些模型确实大量背诵了 Iris、Wine、Titanic 等经典数据集的完整内容。进一步实验表明，模型在"背过"的数据集上表现明显更好，而对数值加入微小扰动后准确率平均下降约 6 个百分点，说明记忆带来了过拟合；相比之下，同样的扰动对模型在全新数据集上的表现几乎没有影响。研究还发现，当去除特征名称、还原为纯统计分类问题时，LLM 的样本效率远不如传统机器学习算法，尤其在特征维度增大时差距更明显，由此推断 LLM 在新数据集上的不错表现主要来自其储存的世界知识而非真正的统计学习能力。这项工作提醒整个社区：在用现有基准评估 LLM 时必须先核查数据集是否已被模型"背过"，作者也为此开源了 tabmemcheck 检测工具包。
原文 arXiv:2404.06209;中英对照 + 大白话阅读 https://aha.fim.ai/paper/2404.06209v3