Data contamination in model evaluation has become increasingly prevalent with the growing popularity of large language models. It allows models to "cheat" via memorisation instead of displaying true capabilities. Therefore, contamination analysis has become an crucial part of reliable model evaluation to validate results. However, existing contamination analysis is usually conducted internally by large language model developers and often lacks transparency and completeness. This paper presents an extensive data contamination report for over 15 popular large language models across six popular multiple-choice QA benchmarks. We also introduce an open-source pipeline that enables the community to perform contamination analysis on customised data and models. Our experiments reveal varying contamination levels ranging from 1\% to 45\% across benchmarks, with the contamination degree increasing rapidly over time. Performance analysis of large language models indicates that data contamination does not necessarily lead to increased model metrics: while significant accuracy boosts of up to 14\% and 7\% are observed on contaminated C-Eval and Hellaswag benchmarks, only a minimal increase is noted on contaminated MMLU. We also find larger models seem able to gain more advantages than smaller models on contaminated test sets.

翻译：数据污染问题在模型评估中随着大语言模型的普及而日益严重。这使得模型能够通过记忆而非展现真实能力来"作弊"。因此，污染分析已成为可靠模型评估中验证结果的关键环节。然而，现有的污染分析通常由大语言模型开发者内部进行，往往缺乏透明度和完整性。本文针对15个以上主流大语言模型在六个通用多选题问答基准上提出了全面的数据污染报告。我们还开源了一套分析流程，使社区能够对自定义数据和模型进行污染分析。实验揭示不同基准的污染程度从1%到45%不等，且污染程度随时间快速增长。大语言模型的性能分析表明，数据污染不一定导致模型指标提升：虽然在受污染的C-Eval和Hellaswag基准上观察到高达14%和7%的显著准确率提升，但在受污染的MMLU上仅发现微小提升。我们还发现，在受污染的测试集上，较大模型似乎比较小模型更能获得优势。