5 papers
Trusted Weights, Treacherous Optimizations? Optimization-Triggered Backdoor Attacks on LLMs
Yifei Wang, Tianlin Li, Xiaohan Zhang +3
Inference optimization is a vital technique for deploying LLMs at scale. Compilation is the most widely adopted optimization technique for LLMs. While it assumes semantic equivalen…
Same Signal, Different Semantics: A Cross-Framework Behavioral Analysis of Software Engineering Agents
Wei Ma, Zhi Chen, Jingxu Gu +3
Behavioral studies of LLM-based software engineering agents extract operational rules about which trajectory shapes correlate with higher resolution rates: that a test step follows…
Hidden Reliability Risks in Large Language Models: Systematic Identification of Precision-Induced Output Disagreements
Yifei Wang, Tianlin Li, Xiaohan Zhang +4
Large language models (LLMs) are increasingly deployed under diverse numerical precision configurations, including standard floating-point formats (e.g., bfloat16 and float16) and…
TokenProber: Jailbreaking Text-to-image Models via Fine-grained Word Impact Analysis
Longtian Wang, Xiaofei Xie, Tianlin Li +2
Text-to-image (T2I) models have significantly advanced in producing high-quality images. However, such models have the ability to generate images containing not-safe-for-work (NSFW…
Benchmarking and Revisiting Code Generation Assessment: A Mutation-Based Approach
Longtian Wang, Tianlin Li, Xiaofei Xie +3
Code Large Language Models (CLLMs) have exhibited outstanding performance in program synthesis, attracting the focus of the research community. The evaluation of CLLM's program syn…