This repository contains the data of the paper "Q&AEval: Benchmarking Secure Coding Ability of LLMs on Real-World Tasks"
The folder "data" contains the questions from StackOverflow, the responses of the LLMs, the extracted and repaired code snippets and the output of the static code analysis tools. Further, it contains the classified programming languages and the questions included in the names datasets (so-small, so-big, so-total).
The folder "cwe" contains information about the filtered CWEs, as well as the manually assigned VUSC CWEs
The folder "code" contains programs written for this paper.
The file prompts.md shows the used prompts. It shows the prompt used for language classification and for querying the LLMs on the SO questions.