SecBench: A Comprehensive Multi-Dimensional Benchmarking Dataset for LLMs in Cybersecurity

2024-12-30

AI Summary

Existing large language models (LLMs) lack a systematic, domain-specific evaluation benchmark for cybersecurity. Method: We introduce SecBench—the first dedicated, multidimensional benchmark for cybersecurity LLM evaluation—covering bilingual (Chinese/English) content, diverse question formats (multiple-choice and short-answer questions), multiple competency dimensions (knowledge retention and logical reasoning), and broad subdomains. It comprises 47,910 high-quality, expert-annotated questions. Our methodology innovates with a “human competition + LLM collaboration” annotation paradigm, an automatically scoring short-answer evaluation agent, and multi-source data fusion with domain-specific knowledge modeling. Contribution/Results: We comprehensively evaluate 13 state-of-the-art LLMs, demonstrating SecBench’s effectiveness and discriminative power. The full dataset and evaluation toolkit are open-sourced, establishing a standardized infrastructure for assessing AI capabilities in cybersecurity.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Application Domains: Security
• Natural Language Processing: (Large) Language Models

Application Category

• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling

Abstract

Evaluating Large Language Models (LLMs) is crucial for understanding their capabilities and limitations across various applications, including natural language processing and code generation. Existing benchmarks like MMLU, C-Eval, and HumanEval assess general LLM performance but lack focus on specific expert domains such as cybersecurity. Previous attempts to create cybersecurity datasets have faced limitations, including insufficient data volume and a reliance on multiple-choice questions (MCQs). To address these gaps, we propose SecBench, a multi-dimensional benchmarking dataset designed to evaluate LLMs in the cybersecurity domain. SecBench includes questions in various formats (MCQs and short-answer questions (SAQs)), at different capability levels (Knowledge Retention and Logical Reasoning), in multiple languages (Chinese and English), and across various sub-domains. The dataset was constructed by collecting high-quality data from open sources and organizing a Cybersecurity Question Design Contest, resulting in 44,823 MCQs and 3,087 SAQs. Particularly, we used the powerful while cost-effective LLMs to (1). label the data and (2). constructing a grading agent for automatic evaluation of SAQs. Benchmarking results on 13 SOTA LLMs demonstrate the usability of SecBench, which is arguably the largest and most comprehensive benchmark dataset for LLMs in cybersecurity. More information about SecBench can be found at our website, and the dataset can be accessed via the artifact link.

Problem

• Cybersecurity
• Language Models
• Evaluation Dataset

Innovation

• SecBench
• Cybersecurity Evaluation
• Large-scale Model Testing