Benchmarking Benchmark Leakage in Large Language Models
This addresses the issue of unfair benchmarking for researchers and developers in the field of large language models, though it is incremental as it builds on existing concerns about data leakage.
The paper tackled the problem of benchmark dataset leakage in large language models, which skews evaluations and causes unfair comparisons, by introducing a detection pipeline using perplexity and n-gram accuracy to analyze 31 models in mathematical reasoning, revealing substantial misuse of test data.
Amid the expanding use of pre-training data, the phenomenon of benchmark dataset leakage has become increasingly prominent, exacerbated by opaque training processes and the often undisclosed inclusion of supervised data in contemporary Large Language Models (LLMs). This issue skews benchmark effectiveness and fosters potentially unfair comparisons, impeding the field's healthy development. To address this, we introduce a detection pipeline utilizing Perplexity and N-gram accuracy, two simple and scalable metrics that gauge a model's prediction precision on benchmark, to identify potential data leakages. By analyzing 31 LLMs under the context of mathematical reasoning, we reveal substantial instances of training even test set misuse, resulting in potentially unfair comparisons. These findings prompt us to offer several recommendations regarding model documentation, benchmark setup, and future evaluations. Notably, we propose the "Benchmark Transparency Card" to encourage clear documentation of benchmark utilization, promoting transparency and healthy developments of LLMs. we have made our leaderboard, pipeline implementation, and model predictions publicly available, fostering future research.