发现大模型生成循环问题并提出实时检测方法
Breaking the Loop: Detecting and Mitigating Denial-of-Service Vulnerabilities in Large Language Models
- 用进化算法黑盒探测大模型的重复生成现象
- 实时检测器准确率达95.24%，F1为0.87
- 适合关注模型安全与性能优化的研究者
大型语言模型（LLMs）在教育、软件开发、医疗、娱乐和法律服务等领域广泛应用。尽管模型可靠性持续提升，但由重复生成引发的延迟问题仍未得到充分研究，可能造成拒绝服务（DoS）漏洞。本文提出RecurentGenerator，一种黑盒进化算法，可高效识别LLama-3和GPT-4o等主流模型中的重复生成场景。同时，我们设计了轻量级实时分类器RecurrentDetector，基于激活模式训练，在检测重复循环任务中达到95.24%准确率和0.87 F1得分。所提方法有效缓解延迟相关漏洞，并公开工具与数据以支持后续研究。
原文摘要 · Abstract (English)
Large Language Models (LLMs) have significantly advanced text understanding and generation, becoming integral to applications across education, software development, healthcare, entertainment, and legal services. Despite considerable progress in improving model reliability, latency remains under-explored, particularly through recurrent generation, where models repeatedly produce similar or identical outputs, causing increased latency and potential Denial-of-Service (DoS) vulnerabilities. We propose RecurrentGenerator, a black-box evolutionary algorithm that efficiently identifies recurrent generation scenarios in prominent LLMs like LLama-3 and GPT-4o. Additionally, we introduce RecurrentDetector, a lightweight real-time classifier trained on activation patterns, achieving 95.24% accuracy and an F1 score of 0.87 in detecting recurrent loops. Our methods provide practical solutions to mitigate latency-related vulnerabilities, and we publicly share our tools and data to support further research.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。