SV-TrustEval-C: Evaluating Structure and Semantic Reasoning in Large Language Models for Source Code Vulnerability Analysis

TL;DR

This paper introduces SV-TrustEval-C, a benchmark for evaluating large language models' abilities to reason about code structure and semantics in C language vulnerability analysis, revealing current models' limitations in logical reasoning.

Contribution

The paper presents a new benchmark, SV-TrustEval-C, specifically designed to assess LLMs' structure and semantic reasoning in source code vulnerability analysis.

Findings

Current LLMs struggle with complex code relationships.

Models rely more on pattern matching than logical reasoning.

Benchmark effectively highlights reasoning deficiencies.

Abstract

As Large Language Models (LLMs) evolve in understanding and generating code, accurately evaluating their reliability in analyzing source code vulnerabilities becomes increasingly vital. While studies have examined LLM capabilities in tasks like vulnerability detection and repair, they often overlook the importance of both structure and semantic reasoning crucial for trustworthy vulnerability analysis. To address this gap, we introduce SV-TrustEval-C, a benchmark designed to evaluate LLMs' abilities for vulnerability analysis of code written in the C programming language through two key dimensions: structure reasoning - assessing how models identify relationships between code elements under varying data and control flow complexities; and semantic reasoning - examining their logical consistency in scenarios where code is structurally and semantically perturbed. Our results show that current LLMs are far from satisfactory in understanding complex code relationships and that their vulnerability analyses rely more on pattern matching than on robust logical reasoning. These findings underscore the effectiveness of the SV-TrustEval-C benchmark and highlight critical areas for enhancing the reasoning capabilities and trustworthiness of LLMs in real-world vulnerability analysis tasks. Our initial benchmark dataset is publicly available.