PaCoST: Paired Confidence Significance Testing for Benchmark Contamination Detection in Large Language Models

TL;DR

This paper introduces PaCoST, a statistical testing method to detect benchmark contamination in large language models, revealing widespread contamination and urging for new evaluation standards.

Contribution

We propose PaCoST, a novel statistical significance testing approach for identifying benchmark data contamination in large language models.

Findings

Most tested models show signs of contamination.

Benchmark contamination affects model performance evaluation.

PaCoST effectively detects contaminated data in LLMs.

Abstract

Large language models (LLMs) are known to be trained on vast amounts of data, which may unintentionally or intentionally include data from commonly used benchmarks. This inclusion can lead to cheatingly high scores on model leaderboards, yet result in disappointing performance in real-world applications. To address this benchmark contamination problem, we first propose a set of requirements that practical contamination detection methods should follow. Following these proposed requirements, we introduce PaCoST, a Paired Confidence Significance Testing to effectively detect benchmark contamination in LLMs. Our method constructs a counterpart for each piece of data with the same distribution, and performs statistical analysis of the corresponding confidence to test whether the model is significantly more confident under the original benchmark. We validate the effectiveness of PaCoST and apply it on popular open-source models and benchmarks. We find that almost all models and benchmarks we tested are suspected contaminated more or less. We finally call for new LLM evaluation methods.