SALLM: Security Assessment of Generated Code

TL;DR

This paper introduces SALLM, a comprehensive framework for benchmarking large language models' ability to generate secure code, addressing gaps in existing datasets and evaluation metrics that overlook security aspects.

Contribution

SALLM provides a new dataset, assessment techniques, and metrics specifically designed to evaluate the security quality of code generated by LLMs.

Findings

SALLM enables systematic benchmarking of secure code generation.

The framework highlights security weaknesses in current LLM-generated code.

It offers tools to improve security-aware code generation by LLMs.

Abstract

With the growing popularity of Large Language Models (LLMs) in software engineers' daily practices, it is important to ensure that the code generated by these tools is not only functionally correct but also free of vulnerabilities. Although LLMs can help developers to be more productive, prior empirical studies have shown that LLMs can generate insecure code. There are two contributing factors to the insecure code generation. First, existing datasets used to evaluate LLMs do not adequately represent genuine software engineering tasks sensitive to security. Instead, they are often based on competitive programming challenges or classroom-type coding tasks. In real-world applications, the code produced is integrated into larger codebases, introducing potential security risks. Second, existing evaluation metrics primarily focus on the functional correctness of the generated code while ignoring security considerations. Therefore, in this paper, we described SALLM, a framework to benchmark LLMs' abilities to generate secure code systematically. This framework has three major components: a novel dataset of security-centric Python prompts, configurable assessment techniques to evaluate the generated code, and novel metrics to evaluate the models' performance from the perspective of secure code generation.