Hello, your paper is highly impressive. While reviewing the source code, I encountered several questions:
For the JSON datasets provided in dataset/MixVul/multi_task, could you clarify the significance of the numbers in the file names? This dataset seems to correspond to the "multi-task learning" described in the paper, but what is the reason for providing multiple files, and what are the specific differences between them?
I am uncertain about how to use the aforementioned multi-task dataset. Could you clarify if the methods starting with VulLLM- in Table 1 of the paper refer to models trained directly using the multi-task dataset?
I hope you can spare some time to address my questions, and I would greatly appreciate any clarification you can provide.
Hello, your paper is highly impressive. While reviewing the source code, I encountered several questions:
dataset/MixVul/multi_task, could you clarify the significance of the numbers in the file names? This dataset seems to correspond to the "multi-task learning" described in the paper, but what is the reason for providing multiple files, and what are the specific differences between them?I hope you can spare some time to address my questions, and I would greatly appreciate any clarification you can provide.