All the experiments are available inside the python notebook (empirical_study_source_code.ipynb). For the modifications made to CodeBert code clone detection pipeline the source code is made available inside Clone-detection-BigCloneBench
The dataset and the BERT model trained for analysing the attention values is available in the following link. Please download the models and store them in the same directory as that of python notebook (empirical_study_source_code.ipynb).
https://drive.google.com/drive/folders/1aAEb1flaSs63EfYHSy7B5_vrYeh4AfW1?usp=sharing
Following repositories are used as reference and modified to write this repository. https://github.com/clarkkev/attention-analysis