19 citations · 51 across the 10 of their papers we have counts for
11 papers
Understanding and Enhancing Backdoor Persistency in LLM Agent Post-Training
Qiusi Zhan, Nian Lyu, Stephanie Ding +5
Developers can build LLM agents by adapting third-party models through benign post-training. We study a supply-chain threat in which an attacker supplies a model with a backdoor: h…
BEAT: Visual Backdoor Attacks on VLM-based Embodied Agents via Contrastive Trigger Learning
Qiusi Zhan, Hyeonjeong Ha, Rui Yang +7
Recent advances in Vision-Language Models (VLMs) have propelled embodied agents by enabling direct perception, reasoning, and planning task-oriented actions from visual inputs. How…
SafeSearch: Do Not Trade Safety for Utility in LLM Search Agents
Qiusi Zhan, Angeline Budiman-Chan, Abdelrahman Zayed +3
Large language model (LLM) based search agents iteratively generate queries, retrieve external information, and reason to answer open-domain questions. While researchers have prima…
Adaptive Attacks Break Defenses Against Indirect Prompt Injection Attacks on LLM Agents
Qiusi Zhan, Richard Fang, Henil Shalin Panchal +1
Large Language Model (LLM) agents exhibit remarkable performance across diverse applications by using external tools to interact with environments. However, integrating external to…
MM-PoisonRAG: Disrupting Multimodal RAG with Local and Global Poisoning Attacks
Hyeonjeong Ha, Qiusi Zhan, Jeonghwan Kim +6
Retrieval-augmented generation (RAG) has become a common practice in multimodal large language models (MLLM) to enhance factual grounding and reduce hallucination. Yet, its relianc…
Voice-Enabled AI Agents can Perform Common Scams
Richard Fang, Dylan Bowman, Daniel Kang
Recent advances in multi-modal, highly capable LLMs have enabled voice-enabled AI agents. These agents are enabling new applications, such as voice-enabled autonomous customer serv…
Teams of LLM Agents can Exploit Zero-Day Vulnerabilities
Yuxuan Zhu, Antony Kellermann, Akul Gupta +4
LLM agents have become increasingly sophisticated, especially in the realm of cybersecurity. Researchers have shown that LLM agents can exploit real-world vulnerabilities when give…
LLM Agents can Autonomously Exploit One-day Vulnerabilities
Richard Fang, Rohan Bindu, Akul Gupta +1
LLMs have becoming increasingly powerful, both in their benign and malicious uses. With the increase in capabilities, researchers have been increasingly interested in their ability…
InjecAgent: Benchmarking Indirect Prompt Injections in Tool-Integrated Large Language Model Agents
Qiusi Zhan, Zhixiang Liang, Zifan Ying +1
Recent work has embodied LLMs as agents, allowing them to access tools, perform actions, and interact with external content (e.g., emails or websites). However, external content in…
LLM Agents can Autonomously Hack Websites
Richard Fang, Rohan Bindu, Akul Gupta +2
In recent years, large language models (LLMs) have become increasingly capable and can now interact with tools (i.e., call functions), read documents, and recursively call themselv…
Removing RLHF Protections in GPT-4 via Fine-Tuning
Qiusi Zhan, Richard Fang, Rohan Bindu +3
As large language models (LLMs) have increased in their capabilities, so does their potential for dual use. To reduce harmful outputs, produces and vendors of LLMs have used reinfo…