The Heap: A Contamination-Free Multilingual Code Dataset for Evaluating Large Language Models
This provides a contamination-free dataset for researchers evaluating large language models on code tasks, though it is incremental as it builds on existing data collection and deduplication methods.
The authors tackled the problem of data contamination in evaluating large language models by releasing The Heap, a large multilingual code dataset covering 57 programming languages that is deduplicated against other open datasets, enabling fair evaluations without significant data cleaning overhead.
The recent rise in the popularity of large language models has spurred the development of extensive code datasets needed to train them. This has left limited code available for collection and use in the downstream investigation of specific behaviors, or evaluation of large language models without suffering from data contamination. To address this problem, we release The Heap, a large multilingual dataset covering 57 programming languages that has been deduplicated with respect to other open datasets of code, enabling researchers to conduct fair evaluations of large language models without significant data cleaning overhead.