Tags → #large language models

Tools
New Automated Vulnerability Detection System Introduced: LProtector
Nov 12, 2024 A new automated vulnerability detection system called LProtector has been introduced, designed for C/C++ codebases and leveraging advanced techniques to improve performance in identifying software vulnerabilities.
Tools
New Benchmark CTIBench Evaluates LLMs in Cyber Threat Intelligence
Nov 12, 2024 A new benchmark called CTIBench has been launched to assess the performance of Large Language Models in Cyber Threat Intelligence tasks through specialized datasets and various evaluation metrics.
Vulnerabilities
New Defense Mechanism Eguard Enhances LLM Embedding Security
Nov 12, 2024 A new defense mechanism called Eguard has been proposed to enhance the security of embedding vector databases used in large language models, significantly reducing the risk of embedding inversion attacks while maintaining performance in various natural language processing tasks.
Tools
PentestAgent Framework Enhances Penetration Testing Automation
Nov 12, 2024 PentestAgent is a new framework that aims to improve penetration testing by integrating large language models and automating various stages of the process.
Vulnerabilities
Smart-LLaMA Method Enhances Smart Contract Vulnerability Detection
Nov 12, 2024 Smart-LLaMA is a new method aimed at improving the detection of vulnerabilities in smart contracts by utilizing large language models and a comprehensive dataset.
Tools
Study Develops LLM-Based Method for Malware Detection
Nov 12, 2024 A recent study from Pennsylvania State University introduces a Large Language Model-based workflow aimed at identifying anti-dynamic analysis techniques used by malware to evade detection in sandbox environments.
Vulnerabilities
Study Examines Vulnerabilities of Language Models to Attacks
Nov 12, 2024 A recent study has identified vulnerabilities in large language models to prompt injection attacks, introducing a novel attack strategy known as Goal-guided Generative Prompt Injection Attack (G2PIA).
Vulnerabilities
Google Advances AI for Software Vulnerability Discovery
Nov 11, 2024 On November 1, Google announced advancements in artificial intelligence for software vulnerability discovery, highlighting the successful identification of a buffer-underflow vulnerability in SQLite by its Big Sleep large language model agent.
Vulnerabilities
Study Examines Vulnerabilities of Large Language Models to Attacks
Nov 11, 2024 A recent study has examined the vulnerabilities of large language models to adversarial attacks, highlighting significant risks associated with cross-model transferability of adversarial prompts.