Haoyang Li
7 indexed papers
Publications per year
Top categories
Frequent co-authors
Research Timeline
This paper proposes a comprehensive taxonomy (SLOT) to systematically categorize security risks, attacks, and defenses specific to Retrieval-Augmented Generation (RAG), clarifying that these risks are distinct from inherent LLM flaws.
Qwen-VLA introduces a unified embodied foundation model that extends vision-language understanding to continuous action generation, enabling robust, multi-task generalization across diverse robotic tasks and embodiments.
Qwen-VLA introduces a unified embodied foundation model that extends vision-language understanding to continuous action generation, enabling robust, multi-task generalization across diverse robotic tasks and embodiments.
ForeMoE is a load balancing system for Mixture-of-Experts reinforcement learning post-training that achieves up to a 1.45x speedup.
Qwen-RobotWorld is a language-conditioned video world model for embodied intelligence that predicts physically grounded future visual trajectories.
This paper studies the robustness of formal explanations under incremental disclosure.
This paper reveals a new vulnerability in Language Model (LLM) inference efficiency caused by persona consistency and proposes RolePlay, a framework to amplify inference costs.
Papers
From Role Prompt to Infinite Thinking: Exploiting Persona Conditioning for Inference Cost Attacks in LLMs
Zhiyi Mou, Wangze Ni, Tianfang Xiao, Haoyang LI +5 more
This paper reveals a new vulnerability in Language Model (LLM) inference efficiency caused by persona consistency and proposes RolePlay, a framework to amplify inference costs.