Computer engineers and programmers have long relied on reverse engineering as a way to copy the functionality of a computer program without copying that program’s copyright-protected code directly.

Researchers have investigated the potential of leveraging pre-trained language models, such as CodeBERT, to enhance source code-related tasks. Previous methodologies have relied on CodeBERT’s ‘[CLS]’ ...