Beko

The 'Beko' dataset contains 2,604 bot-generated pull request comments labeled by software engineers as fixed or won't fix, and it is used to evaluate the effectiveness of automated code review bots in an industrial setting.