6 papers
PipeWeave: Synergizing Analytical and Learning Models for Unified GPU Performance Prediction
Kaixuan Zhang, Yunfan Cui, Shuhao Zhang +8
The rapid expansion of Transformer-based large language models has dramatically increased the need for high-performance GPUs. As a result, there is growing demand for fast, accurat…
MAD-Spear: A Conformity-Driven Prompt Injection Attack on Multi-Agent Debate Systems
Yu Cui, Hongyang Du
Multi-agent debate (MAD) systems leverage collaborative interactions among large language models (LLMs) agents to improve reasoning capabilities. While recent studies have focused…
Practical Reasoning Interruption Attacks on Reasoning Large Language Models
Yu Cui, Cong Zuo
Reasoning large language models (RLLMs) have demonstrated outstanding performance across a variety of tasks, yet they also expose numerous security vulnerabilities. Most of these v…
Token-Efficient Prompt Injection Attack: Provoking Cessation in LLM Reasoning via Adaptive Token Compression
Yu Cui, Yujun Cai, Yiwei Wang
While reasoning large language models (LLMs) demonstrate remarkable performance across various tasks, they also contain notable security vulnerabilities. Recent research has uncove…
Process or Result? Manipulated Ending Tokens Can Mislead Reasoning LLMs to Ignore the Correct Reasoning Steps
Yu Cui, Bryan Hooi, Yujun Cai +1
Recent reasoning large language models (LLMs) have demonstrated remarkable improvements in mathematical reasoning capabilities through long Chain-of-Thought. The reasoning tokens o…
PPLqa: An Unsupervised Information-Theoretic Quality Metric for Comparing Generative Large Language Models
Gerald Friedland, Xin Huang, Yueying Cui +3
We propose PPLqa, an easy to compute, language independent, information-theoretic metric to measure the quality of responses of generative Large Language Models (LLMs) in an unsupe…