# Evaluation data contamination in LLMs: how do we measure it and (when) does it matter?

arXiv: https://arxiv.org/abs/2411.03923 · HTML: https://mlsanity.com/paper/2411.03923/

## Key facts

- Cited by 32 papers in the map (data through 14 Oct 2026).
- Builds on 28 of its 40 references that are in the map.
- First posted on arXiv 6 Nov 2024.

## Builds on (most cited first, 28 of 28)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Measuring Mathematical Problem Solving With the MATH Dataset](https://mlsanity.com/paper/2103.03874/) (2021)
- [PaLM: Scaling Language Modeling with Pathways](https://mlsanity.com/paper/2204.02311/) (2022)
- [HellaSwag: Can a Machine Really Finish Your Sentence?](https://mlsanity.com/paper/1905.07830/) (2019)
- [Finetuned Language Models Are Zero-Shot Learners](https://mlsanity.com/paper/2109.01652/) (2021)
- [Program Synthesis with Large Language Models](https://mlsanity.com/paper/2108.07732/) (2021)
- [TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension](https://mlsanity.com/paper/1705.03551/) (2017)
- [PIQA: Reasoning about Physical Commonsense in Natural Language](https://mlsanity.com/paper/1911.11641/) (2019)
- [Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models](https://mlsanity.com/paper/2206.04615/) (2022)
- [Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Them](https://mlsanity.com/paper/2210.09261/) (2022)
- [Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling](https://mlsanity.com/paper/2304.01373/) (2023)
- [The FineWeb Datasets: Decanting the Web for the Finest Text Data at Scale](https://mlsanity.com/paper/2406.17557/) (2024)
- [GLaM: Efficient Scaling of Language Models with Mixture-of-Experts](https://mlsanity.com/paper/2112.06905/) (2021)
- [Dolma: an Open Corpus of Three Trillion Tokens for Language Model Pretraining Research](https://mlsanity.com/paper/2402.00159/) (2024)
- [NLP Evaluation in trouble: On the Need to Measure LLM Data Contamination for each Benchmark](https://mlsanity.com/paper/2310.18018/) (2023)
- [Proving Test Set Contamination in Black Box Language Models](https://mlsanity.com/paper/2310.17623/) (2023)
- [Leak, Cheat, Repeat: Data Contamination and Evaluation Malpractices in Closed-Source LLMs](https://mlsanity.com/paper/2402.03927/) (2024)
- [Rethinking Benchmark and Contamination for Language Models with Rephrased Samples](https://mlsanity.com/paper/2311.04850/) (2023)
- [Quantifying Privacy Risks of Masked Language Models Using Membership Inference Attacks](https://mlsanity.com/paper/2203.03929/) (2022)
- [LAB-Bench: Measuring Capabilities of Language Models for Biology Research](https://mlsanity.com/paper/2407.10362/) (2024)
- [Stop Uploading Test Data in Plain Text: Practical Strategies for Mitigating Data Contamination by Evaluation Benchmarks](https://mlsanity.com/paper/2305.10160/) (2023)
- [What's In My Big Data?](https://mlsanity.com/paper/2310.20707/) (2023)
- [Investigating Data Contamination for Pre-training Language Models](https://mlsanity.com/paper/2401.06059/) (2024)
- [WorldSense: A Synthetic Benchmark for Grounded Reasoning in Large Language Models](https://mlsanity.com/paper/2311.15930/) (2023)
- [Data Contamination Report from the 2024 CONDA Shared Task](https://mlsanity.com/paper/2407.21530/) (2024)

## Built on by (most cited first, 30 of 32)

- [Tulu 3: Pushing Frontiers in Open Language Model Post-Training](https://mlsanity.com/paper/2411.15124/) (2024)
- [Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2504.12501/) (2025)
- [CodeI/O: Condensing Reasoning Patterns via Code Input-Output Prediction](https://mlsanity.com/paper/2502.07316/) (2025)
- [HARP: A challenging human-annotated math reasoning benchmark](https://mlsanity.com/paper/2412.08819/) (2024)
- [Language Models May Verbatim Complete Text They Were Not Explicitly Trained On](https://mlsanity.com/paper/2503.17514/) (2025)
- [Overestimation in LLM Evaluation: A Controlled Large-Scale Study on Data Contamination's Impact on Machine Translation](https://mlsanity.com/paper/2501.18771/) (2025)
- [LLM Knowledge is Brittle: Truthfulness Representations Rely on Superficial Resemblance](https://mlsanity.com/paper/2510.11905/) (2025)
- [STAMP Your Content: Proving Dataset Membership via Watermarked Rephrasings](https://mlsanity.com/paper/2504.13416/) (2025)
- [Detecting Data Contamination in LLMs via In-Context Learning](https://mlsanity.com/paper/2510.27055/) (2025)
- [Detecting Benchmark Contamination Through Watermarking](https://mlsanity.com/paper/2502.17259/) (2025)
- [Evaluating the Robustness of Large Language Model Safety Guardrails Against Adversarial Attacks](https://mlsanity.com/paper/2511.22047/) (2025)
- [The Impact of Post-training on Data Contamination](https://mlsanity.com/paper/2601.06103/) (2026)
- [Scaling Large Reasoning Models beyond Human Supervision: A Path toward Superintelligence](https://mlsanity.com/paper/2608.31075/) (2026)
- [Data-Centric Lessons To Improve Speech-Language Pretraining](https://mlsanity.com/paper/2510.20860/) (2025)
- [Theory-Grounded Evaluation of Human-Like Fallacy Patterns in LLM Reasoning](https://mlsanity.com/paper/2506.11128/) (2025)
- [LLM Benchmark Datasets Should Be Contamination-Resistant](https://mlsanity.com/paper/2605.19999/) (2026)
- [Temporal Leakage in LLM Backtesting: Measurement, Validation, and Adjusted Scores](https://mlsanity.com/paper/2608.02985/) (2026)
- [Two Causally Related Needles in a Video Haystack](https://mlsanity.com/paper/2505.19853/) (2025)
- [Obscuring Data Contamination Through Translation: Evidence from Arabic Corpora](https://mlsanity.com/paper/2601.14994/) (2026)
- [Benchmarking at the Edge of Comprehension](https://mlsanity.com/paper/2602.14307/) (2026)
- [QuickScope: Certifying Hard Questions in Dynamic LLM Benchmarks](https://mlsanity.com/paper/2604.17842/) (2026)
- [Measuring AI Reasoning: A Guide for Researchers](https://mlsanity.com/paper/2605.02442/) (2026)
- [Same Question, Different Answers: Evaluating LLM Reliability Beyond Accuracy](https://mlsanity.com/paper/2607.22554/) (2026)
- [LLMs for Resource Allocation: A Participatory Budgeting Approach to Inferring Preferences](https://mlsanity.com/paper/2508.06060/) (2025)
- [How Much Reasoning Do Retrieval-Augmented Models Add beyond LLMs? A Benchmarking Framework for Multi-Hop Inference over Hybrid Knowledge](https://mlsanity.com/paper/2602.10210/) (2026)
- [Learning to Detect Language Model Training Data via Active Reconstruction](https://mlsanity.com/paper/2602.19020/) (2026)
- [Are Large Language Models Truly Smarter Than Humans?](https://mlsanity.com/paper/2603.16197/) (2026)
- [Physics-R1: An Audited Olympiad Corpus and Released Verifiers for Visual Physics Reasoning](https://mlsanity.com/paper/2605.14040/) (2026)
- [Anchored or Drifting: What Recursive Self-Generation Reveals About Training Data](https://mlsanity.com/paper/2606.31991/) (2026)
- [LeakScale: Estimating the Causal Effect of Benchmark Exposure](https://mlsanity.com/paper/2609.27176/) (2026)

## Abstract (arXiv)

Hampering the interpretation of benchmark scores, evaluation data contamination has become a growing concern in the evaluation of LLMs, and an active area of research studies its effects. While evaluation data contamination is easily understood intuitively, it is surprisingly difficult to define precisely which samples should be considered contaminated and, consequently, how it impacts benchmark scores. We propose that these questions should be addressed together and that contamination metrics can be assessed based on whether models benefit from the examples they mark contaminated. We propose a novel analysis method called ConTAM, and show with a large scale survey of existing and novel n-gram based contamination metrics across 13 benchmarks and 7 models from 2 different families that ConTAM can be used to better understand evaluation data contamination and its effects. We find that contamination may have a much larger effect than reported in recent LLM releases and benefits models differently at different scales. We also find that considering only the longest contaminated substring provides a better signal than considering a union of all contaminated substrings, and that doing model and benchmark specific threshold analysis greatly increases the specificity of the results. Lastly, we investigate the impact of hyperparameter choices, finding that, among other things, both using larger values of n and disregarding matches that are infrequent in the pre-training data lead to many false negatives. With ConTAM, we provide a method to empirically ground evaluation data contamination metrics in downstream effects. With our exploration, we shed light on how evaluation data contamination can impact LLMs and provide insight into the considerations important when doing contamination analysis. We end our paper by discussing these in more detail and providing concrete suggestions for future work.