arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2311.04850cs.CLcs.AI
收藏

重新思考语言模型基准测试与样本改写导致的污染

Rethinking Benchmark and Contamination for Language Models with Rephrased Samples

UC Berkeley(加州大学伯克利分校);Shanghai Jiao Tong University(上海交通大学)

Shuo Yang, Wei-Lin Chiang, Lianmin Zheng, Joseph E. Gonzalez, Ion Stoica

AI总结:

针对语言模型基准测试污染问题,本文揭示字符串匹配去污的不足,提出基于LLM的更强去污方法,并发现广泛数据集存在显著重叠,呼吁采用更强去污与全新考试评估模型。

摘要:

大型语言模型越来越多地在人类产生的所有数据上进行训练。许多人担心公共基准测试的可信度,因为预训练或微调数据集中可能存在污染。虽然大多数数据去污工作采用字符串匹配(如n-gram重叠)来移除基准数据,但我们表明这些方法是不够的,测试数据的简单变体(如改写、翻译)可以轻易绕过这些去污措施。此外,我们证明,如果不清除这类测试数据变体,一个13B模型可以轻易过拟合测试基准,并获得与GPT-4相当的超高表现。我们在广泛使用的基准测试如MMLU、GSM8k和HumanEval中验证了这些观察。为应对这一日益增长的风险,我们提出了一种更强的基于LLM的去污方法,并将其应用于广泛使用的预训练和微调数据集,揭示了大量此前未知的测试重叠。例如,在RedPajama-Data-1T和StarCoder-Data等预训练集中,我们识别出8-18%的HumanEval基准重叠。有趣的是,我们还发现GPT-3.5/4生成的合成数据集中也存在此类污染,表明存在无意污染的潜在风险。我们敦促社区在使用公共基准测试时采用更强的去污方法。此外,我们呼吁社区积极开发全新的一次性考试来准确评估模型。我们的去污工具公开可用,网址为https://github.com/lm-sys/llm-decontaminator。

英文摘要:

Large language models are increasingly trained on all the data ever produced by humans. Many have raised concerns about the trustworthiness of public benchmarks due to potential contamination in pre-training or fine-tuning datasets. While most data decontamination efforts apply string matching (e.g., n-gram overlap) to remove benchmark data, we show that these methods are insufficient, and simple variations of test data (e.g., paraphrasing, translation) can easily bypass these decontamination measures. Furthermore, we demonstrate that if such variation of test data is not eliminated, a 13B model can easily overfit a test benchmark and achieve drastically high performance, on par with GPT-4. We validate such observations in widely used benchmarks such as MMLU, GSK8k, and HumanEval. To address this growing risk, we propose a stronger LLM-based decontamination method and apply it to widely used pre-training and fine-tuning datasets, revealing significant previously unknown test overlap. For example, in pre-training sets such as RedPajama-Data-1T and StarCoder-Data, we identified that 8-18\% of the HumanEval benchmark overlaps. Interestingly, we also find such contamination in synthetic dataset generated by GPT-3.5/4, suggesting a potential risk of unintentional contamination. We urge the community to adopt stronger decontamination approaches when using public benchmarks. Moreover, we call for the community to actively develop fresh one-time exams to evaluate models accurately. Our decontamination tool is publicly available at https://github.com/lm-sys/llm-decontaminator.

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。

↑