LastingBench: Defend Benchmarks Against Knowledge Leakage
This addresses the issue of benchmark reliability for AI researchers and developers, though it is incremental as it builds on prior leakage detection work.
The paper tackles the problem of large language models 'cheating' on QA benchmarks through data memorization, which undermines evaluation validity, and introduces LastingBench to mitigate this by rewriting leakage points to counterfactual ones, showing significant performance gaps in evaluations.
The increasing complexity of large language models (LLMs) raises concerns about their ability to "cheat" on standard Question Answering (QA) benchmarks by memorizing task-specific data. This undermines the validity of benchmark evaluations, as they no longer reflect genuine model capabilities but instead the effects of data leakage. While prior work has focused on detecting such leakage, little attention has been given to mitigating its impact and preserving the long-term utility of benchmarks. In this paper, we introduce LastingBench, a novel framework designed to continuously reinforce and safeguard existing benchmarks against knowledge leakage. LastingBench identifies leakage points in the context through perturbation, then rewrites the leakage points to counterfactual ones-disrupting memorization while preserving the benchmark's original evaluative intent. Evaluations of state-of-the-art QA benchmarks show significant performance gaps, highlighting the efficacy of LastingBench in reducing memorization effects. LastingBench offers a practical and scalable solution to ensure benchmark robustness over time, promoting fairer and more interpretable evaluations of LLMs.