搜索时数据污染
Search-Time Data Contamination
AI总结：
本文提出搜索时污染（STC）问题，揭示基于搜索的LLM智能体在评估中会从HuggingFace等在线平台检索到含答案的测试数据，并通过实验量化其影响，进而提出基准设计与结果报告的最佳实践。
摘要：
数据污染是指评估数据泄漏到模型训练数据中，导致对原本应留出的测试集产生过拟合，从而损害测试有效性。我们识别出一个类似的问题，即搜索时污染（STC），它出现在评估基于搜索的LLM智能体时——这些智能体在回答用户查询时使用工具从在线来源收集信息。当检索步骤浮出一个包含测试问题（或近似重复项）及其答案的来源时，就会发生STC，使智能体能够复制答案而不是真正进行推断或推理，从而破坏基准的完整性。我们发现HuggingFace这一托管评估数据集的在线平台出现在基于搜索的智能体日志所检索到的来源中。因此，智能体经常在其推理链中明确承认从HuggingFace发现了问题-答案对。在三个常用的能力基准上：Humanity's Last Exam（HLE）、SimpleQA和GPQA，我们证明对于约3%的问题，基于搜索的智能体会直接在HuggingFace上找到带有真实标签的数据集。当数百万个评估查询针对同一基准时，即使是小规模、重复的泄漏也可能加速该基准的过时，缩短其预期的生命周期。在屏蔽HuggingFace之后，我们观察到受污染子集上的准确率下降了约15%。我们进一步通过消融实验表明，HuggingFace上公开可访问的评估数据集可能不是STC的唯一来源。为此，我们在结论中提出了基准设计和结果报告的最佳实践，以应对这种新型泄漏，并确保对基于搜索的LLM智能体进行可信的评估。为了便于审计评估结果，我们还公开发布了实验的完整日志。
英文摘要：
Data contamination refers to the leakage of evaluation data into model training data, resulting in overfitting to supposedly held-out test sets and compromising test validity. We identify an analogous issue, search-time contamination (STC), in evaluating search-based LLM agents which use tools to gather information from online sources when answering user queries. STC occurs when the retrieval step surfaces a source containing the test question (or a near-duplicate) alongside its answer, enabling agents to copy rather than genuinely infer or reason, undermining benchmark integrity. We find that HuggingFace, an online platform hosting evaluation datasets, appears among retrieved sources in search based agent logs. Consequently, agents often explicitly acknowledge discovering question answer pairs from HuggingFace within their reasoning chains. On three commonly used capability benchmarks: Humanity's Last Exam (HLE), SimpleQA, and GPQA, we demonstrate that for approximately 3% of questions, search-based agents directly find the datasets with ground truth labels on HuggingFace. When millions of evaluation queries target the same benchmark, even small, repeated leaks can accelerate the benchmark's obsolescence, shortening its intended lifecycle. After HuggingFace is blocked, we observe a drop in accuracy on the contaminated subset of approximately 15%. We further show through ablation experiments that publicly accessible evaluation datasets on HuggingFace may not be the sole source of STC. To this end, we conclude by proposing best practices for benchmark design and result reporting to address this novel form of leakage and ensure trustworthy evaluation of search-based LLM agents. To facilitate the auditing of evaluation results, we also publicly release the complete logs from our experiments.
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。
