cs.AI（2025-09-06）

📊 共 9 篇论文 | 🔗 1 篇有代码

🎯 兴趣领域导航

#	题目	一句话要点	标签
1	OccVLA: Vision-Language-Action Model with Implicit 3D Occupancy Supervision	OccVLA：利用隐式3D Occupancy监督的视觉-语言-动作模型	vision-language-action large language model multimodal
2	Towards Meta-Cognitive Knowledge Editing for Multimodal LLMs	提出MIND框架，增强多模态LLM的元认知知识编辑能力，解决现有方法缺乏深层认知的问题。	large language model multimodal
3	Reasoning Introduces New Poisoning Attacks Yet Makes Them More Complicated	针对具备推理能力的LLM，提出分解推理中毒攻击，但模型展现出一定鲁棒性	large language model chain-of-thought
4	Decoding Latent Attack Surfaces in LLMs: Prompt Injection via HTML in Web Summarization	通过HTML隐蔽注入攻击揭示LLM的脆弱性	large language model
5	Red-Teaming Coding Agents from a Tool-Invocation Perspective: An Empirical Security Assessment	针对代码生成Agent工具调用环节的安全风险，提出ToolLeak漏洞和双通道注入攻击。	large language model
6	EchoLeak: The First Real-World Zero-Click Prompt Injection Exploit in a Production LLM System	EchoLeak：首个在生产LLM系统中实现的零点击Prompt注入漏洞利用	large language model

#	题目	一句话要点	标签	🔗
7	Hyperbolic Large Language Models	提出基于双曲几何的大语言模型(HypLLMs)，增强语义表示学习和多尺度推理。	representation learning large language model	✅
8	MSRFormer: Road Network Representation Learning using Multi-scale Feature Fusion of Heterogeneous Spatial Interactions	MSRFormer：利用异构空间交互的多尺度特征融合进行道路网络表征学习	representation learning contrastive learning
9	Chatbot To Help Patients Understand Their Health	提出NoteAid-Chatbot，利用多智能体LLM和强化学习提升患者健康知识理解	reinforcement learning PPO large language model