Coffee
Emerging3papers using it
2023first seen
The 'Coffee' dataset is a collected resource specifically designed for evaluating code fixing with feedback, aimed at improving the performance of open-source code LLMs in generating helpful and accurate feedback for code editing.