RN-F: A Novel Approach for Mitigating Contaminated Data in Large Language Models

TL;DR

This paper introduces RN-F, a lightweight, model-agnostic method for detecting contaminated data in large language models, significantly improving detection accuracy without extra computational overhead.

Contribution

We propose RN-F, a novel residual-noise fingerprinting framework that effectively identifies contaminated data in LLMs without additional floating-point operations.

Findings

RN-F outperforms existing methods by up to 10.5% in detection metrics.

RN-F is lightweight, gradient-free, and model-agnostic.

The approach is effective across multiple LLMs and datasets.

Abstract

Large Language Models (LLMs) have become foundational in modern artificial intelligence, powering a wide range of applications from code generation and virtual assistants to scientific research and enterprise automation. However, concerns about data contamination--where test data overlaps with training data--have raised serious questions about the reliability of these applications. Despite awareness of this issue, existing methods fall short in effectively identifying or mitigating contamination. In this paper, we propose Residual-Noise Fingerprinting (RN-F), a novel framework for detecting contaminated data in LLMs. RN-F is a single-pass, gradient-free detection method that leverages residual signal patterns without introducing additional floating-point operations. Our approach is lightweight, model-agnostic, and efficient. We evaluate RN-F on multiple LLMs across various contaminated datasets and show that it consistently outperforms existing state-of-the-art methods, achieving performance improvements of up to 10.5% in contamination detection metrics.