Detecting Data Contamination in LLMs via In-Context Learning
This addresses the issue of data contamination for researchers and practitioners evaluating LLMs, though it is incremental as it builds on existing in-context learning techniques.
The paper tackles the problem of detecting training data contamination in large language models by introducing CoDeC, a method that uses in-context learning to measure contamination, resulting in interpretable scores that clearly separate seen and unseen datasets and reveal memorization in models with undisclosed training data.
We present Contamination Detection via Context (CoDeC), a practical and accurate method to detect and quantify training data contamination in large language models. CoDeC distinguishes between data memorized during training and data outside the training distribution by measuring how in-context learning affects model performance. We find that in-context examples typically boost confidence for unseen datasets but may reduce it when the dataset was part of training, due to disrupted memorization patterns. Experiments show that CoDeC produces interpretable contamination scores that clearly separate seen and unseen datasets, and reveals strong evidence of memorization in open-weight models with undisclosed training corpora. The method is simple, automated, and both model- and dataset-agnostic, making it easy to integrate with benchmark evaluations.