Search-Time Data Contamination

• URL: http://arxiv.org/abs/2508.13180v1
• Date: Tue, 12 Aug 2025 22:52:21 GMT
• Title: Search-Time Data Contamination
• Abstract summary: We identify an analogous issue, search-time contamination (STC), in evaluating search-based LLM agents.<n>We find that HuggingFace, an online platform hosting evaluation datasets, appears among retrieved sources in search based agent logs.<n>We propose best practices for benchmark design and result reporting to address this novel form of leakage.
• Score: 18.94571261664399
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: 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.

Related papers

• Search-Time Contamination in Deep Research Agents: Measuring Performance Inflation in Public Benchmark Evaluation [17.913171310675306]
  We study Search-Time Contamination (STC) in deep research agent evaluation.<n>We find that STC is widespread and can inflate performance by up to 4%.<n>We advocate contamination-aware practices, including isolated sandboxes, transparent search trajectories, and controlled benchmark access.
  arXiv  Detail & Related papers  (2026-06-03T07:11:36Z)
• LiveBrowseComp: Are Search Agents Searching, or Just Verifying What They Already Know? [32.434901767447165]
  LLM-based search agents rely on intrinsic knowledge rather than external evidence.<n>LiveBrowseComp is a benchmark designed to evaluate agents beyond intrinsic coverage.<n>All evaluated agents fall below 2% closed-book accuracy, search-augmented scores drop by 25-40 points relative to BrowseComp, and prior model rankings no longer reliably predict performance.
  arXiv  Detail & Related papers  (2026-05-27T16:39:57Z)
• iAgentBench: Benchmarking Sensemaking Capabilities of Information-Seeking Agents on High-Traffic Topics [9.25340189071758]
  We present iAgentBench, a dynamic ODQA benchmark for cross-source sensemaking.<n>iAgentBench draws seed topics from real-world attention signals and uses common user intent patterns to construct user-like questions.<n>Each instance is released with traceable evidence and auditable intermediate artifacts that support contamination checks.
  arXiv  Detail & Related papers  (2026-03-04T22:40:08Z)
• SAGE: Steerable Agentic Data Generation for Deep Search with Execution Feedback [68.60326181052658]
  We propose an agentic pipeline that automatically generates high quality, difficulty-controlled deep search question-answer pairs.<n>Our pipeline, SAGE, consists of a data generator which proposes QA pairs and a search agent which attempts to solve the generated question.<n>Our intrinsic evaluation shows SAGE generates questions that require diverse reasoning strategies, while significantly increases the correctness and difficulty of the generated data.
  arXiv  Detail & Related papers  (2026-01-26T06:37:56Z)
• Benchmarking Deep Search over Heterogeneous Enterprise Data [73.55304268238474]
  We present a new benchmark for evaluating a form of retrieval-augmented generation (RAG)<n>RAG requires source-aware, multi-hop reasoning over diverse, sparsed, but related sources.<n>We build it using a synthetic data pipeline that simulates business across product planning, development, and support stages.
  arXiv  Detail & Related papers  (2025-06-29T08:34:59Z)
• Maximally-Informative Retrieval for State Space Model Generation [59.954191072042526]
  We introduce Retrieval In-Context Optimization (RICO) to minimize model uncertainty for a particular query at test-time.<n>Unlike traditional retrieval-augmented generation (RAG), which relies on externals for document retrieval, our approach leverages direct feedback from the model.<n>We show that standard top-$k$ retrieval with model gradients can approximate our optimization procedure, and provide connections to the leave-one-out loss.
  arXiv  Detail & Related papers  (2025-06-13T18:08:54Z)
• Training on the Benchmark Is Not All You Need [52.01920740114261]
  We propose a simple and effective data leakage detection method based on the contents of multiple-choice options.<n>Our method is able to work under gray-box conditions without access to model training data or weights.<n>We evaluate the degree of data leakage of 35 mainstream open-source LLMs on four benchmark datasets.
  arXiv  Detail & Related papers  (2024-09-03T11:09:44Z)
• Investigating Data Contamination in Modern Benchmarks for Large Language Models [27.479260572913724]
  Recent observations have underscored a disparity between the inflated benchmark scores and the actual performance of LLMs. We study data contamination by proposing two methods tailored for both open-source and proprietary LLMs. We find that certain commercial LLMs could surprisingly guess the missing option in various test sets.
  arXiv  Detail & Related papers  (2023-11-16T11:03:04Z)
• Rethinking Benchmark and Contamination for Language Models with Rephrased Samples [49.18977581962162]
  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.
  arXiv  Detail & Related papers  (2023-11-08T17:35:20Z)
• NLP Evaluation in trouble: On the Need to Measure LLM Data Contamination for each Benchmark [19.875954121100005]
  We argue that the classical evaluation on Natural Language Processing (NLP) tasks using annotated benchmarks is in trouble. The worst kind of data contamination happens when a Large Language Model (LLM) is trained on the test split of a benchmark, and then evaluated in the same benchmark. This position paper defines different levels of data contamination and argues for a community effort.
  arXiv  Detail & Related papers  (2023-10-27T09:48:29Z)
• Detecting Pretraining Data from Large Language Models [90.12037980837738]
  We study the pretraining data detection problem. Given a piece of text and black-box access to an LLM without knowing the pretraining data, can we determine if the model was trained on the provided text? We introduce a new detection method Min-K% Prob based on a simple hypothesis.
  arXiv  Detail & Related papers  (2023-10-25T17:21:23Z)
• Stop Uploading Test Data in Plain Text: Practical Strategies for Mitigating Data Contamination by Evaluation Benchmarks [70.39633252935445]
  Data contamination has become prevalent and challenging with the rise of models pretrained on large automatically-crawled corpora. For closed models, the training data becomes a trade secret, and even for open models, it is not trivial to detect contamination. We propose three strategies that can make a difference: (1) Test data made public should be encrypted with a public key and licensed to disallow derivative distribution; (2) demand training exclusion controls from closed API holders, and protect your test data by refusing to evaluate without them; and (3) avoid data which appears with its solution on the internet, and release the web-page context of internet-derived
  arXiv  Detail & Related papers  (2023-05-17T12:23:38Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.