6 accepted papers
Backdoor attacks embed malicious behaviors into Large Language Models (LLMs), enabling adversaries to trigger harmful outputs or bypass safety controls. However, the persistence of the implanted backdoors under user-driven post-deployment continual fine-tuning has been rarely examined. Most prior wo
Intelligent agents powered by large language models (LLMs) have recently demonstrated impressive capabilities and gained increasing popularity on social media platforms. While LLM agents are reshaping the ecology of social media, there exists a current gap in conducting a comprehensive evaluation of
The dimensions of components have recently expanded in range from the micrometer to centimeter scale in MEMS assembly, necessitating the regulation of adhesion force across a broad range to accommodate cross-scale micromanipulation tasks. Inspired by gecko, anisotropic microwedges can effectively re
Autonomous robotic assembly of flexible printed circuit (FPC) for electronic devices requires accurate characterization of contact states between terminals and slots. This task becomes particularly challenging in scenarios with limited visibility and high precision demands. While visual methods ofte
This study presents a soft gripper incorporating a novel belt loop actuated adhesion design, offering a comprehensive solution for the fabrication, variable-scale actuation, and contact sensing of directional adhesives. The solution facilitates high-resolution control of directional adhesives with r
Backdoor attacks in reinforcement learning (RL) have previously employed intense attack strategies to ensure attack success. However, these methods suffer from high attack costs and increased detectability. In this work, we propose a novel approach, BadRL, which focuses on conducting highly sparse b…