We will publish the dataset and code once the paper is accepted.
Our data collection process, akin to CVEFixes, involves gathering CVE records from 2022 to October 2023. Following deduplication with CVEFixes and Big-Vul, our collected dataset encompasses 2,701 CVE records spanning 1,136 open-source projects, featuring 73 CWE IDs.