README.md

metadata

license: apache-2.0
task_categories:
  - text-generation
language:
  - en
size_categories:
  - n<1K

SALLM: Security Assessment of Generated Code

This dataset contains 100 samples to evaluate the security of the LLM generated code.

Dataset Details

Dataset Description

• Curated by: Mohammed Latif Siddiq, Joanna C. S. Santos, Sajith Devareddy, Anna Muller

Dataset Sources [optional]

• Repository: https://github.com/s2e-lab/SALLM
• Paper: https://dl.acm.org/doi/abs/10.1145/3691621.3694934

Dataset Structure

• id: The unique identifier of the prompt
• technique: The security technique that the prompt is based on
• source: The source of the prompt
• prompt: The prompt itself
• insecure_code: The insecure code snippet

Citation [optional]

BibTeX:

@inproceedings{siddiq2024sallm,
  author={Siddiq, Mohammed Latif and Santos, Joanna C. S. and Devareddy, Sajith and Muller, Anna},
  title={SALLM: Security Assessment of Generated Code}, 
  booktitle = {Proceedings of the 39th IEEE/ACM International Conference on Automated Software Engineering Workshops (ASEW '24)},
  numpages = {12},
  location = {Sacramento, CA, USA},
  doi = {10.1145/3691621.3694934}
  series = {ASEW '24}
}

Dataset Card Authors

Mohammed Latif Siddiq

Dataset Card Contact

Website: https://lsiddiqsunny.github.io/