arXiv subjects

Zhun Wang

Publications and source records attributed to Zhun Wang.

27 records · Page 2Linked to original sources

arXiv2025

PromptArmor: Simple yet Effective Prompt Injection Defenses

Despite their potential, recent research has demonstrated that LLM agents are vulnerable to prompt injection attacks, where malicious prompts are injected into the agent's input, causing it to perform an attacker-specified task rather than the intended task provided by the user. In this paper, we present PromptArmor, a simple yet effective defense against prompt injection attacks. Specifically, PromptArmor prompts an off-the-shelf LLM to detect and remove potential injected prompts from the input before the agent processes it. Our results show that PromptArmor can accurately identify and remove injected prompts. For example, using GPT-4o, GPT-4.1, or o4-mini, PromptArmor achieves both a false positive rate and a false negative rate below 1% on the AgentDojo benchmark. Moreover, after removing injected prompts with PromptArmor, the attack success rate drops to below 1%. We also demonstrate PromptArmor's effectiveness against adaptive attacks and explore different strategies for prompting an LLM. We recommend that PromptArmor be adopted as a standard baseline for evaluating new defenses against prompt injection attacks.

cs.CR↗

arXiv2025

AgentVigil: Generic Black-Box Red-teaming for Indirect Prompt Injection against LLM Agents

The strong planning and reasoning capabilities of Large Language Models (LLMs) have fostered the development of agent-based systems capable of leveraging external tools and interacting with increasingly complex environments. However, these powerful features also introduce a critical security risk: indirect prompt injection, a sophisticated attack vector that compromises the core of these agents, the LLM, by manipulating contextual information rather than direct user prompts. In this work, we propose a generic black-box fuzzing framework, AgentVigil, designed to automatically discover and exploit indirect prompt injection vulnerabilities across diverse LLM agents. Our approach starts by constructing a high-quality initial seed corpus, then employs a seed selection algorithm based on Monte Carlo Tree Search (MCTS) to iteratively refine inputs, thereby maximizing the likelihood of uncovering agent weaknesses. We evaluate AgentVigil on two public benchmarks, AgentDojo and VWA-adv, where it achieves 71% and 70% success rates against agents based on o3-mini and GPT-4o, respectively, nearly doubling the performance of baseline attacks. Moreover, AgentVigil exhibits strong transferability across unseen tasks and internal LLMs, as well as promising results against defenses. Beyond benchmark evaluations, we apply our attacks in real-world environments, successfully misleading agents to navigate to arbitrary URLs, including malicious sites.

cs.CR↗

arXiv2025

COSMIC: Generalized Refusal Direction Identification in LLM Activations

Large Language Models (LLMs) encode behaviors such as refusal within their activation space, yet identifying these behaviors remains a significant challenge. Existing methods often rely on predefined refusal templates detectable in output tokens or require manual analysis. We introduce \textbf{COSMIC} (Cosine Similarity Metrics for Inversion of Concepts), an automated framework for direction selection that identifies viable steering directions and target layers using cosine similarity - entirely independent of model outputs. COSMIC achieves steering performance comparable to prior methods without requiring assumptions about a model's refusal behavior, such as the presence of specific refusal tokens. It reliably identifies refusal directions in adversarial settings and weakly aligned models, and is capable of steering such models toward safer behavior with minimal increase in false refusals, demonstrating robustness across a wide range of alignment conditions.

cs.CL↗

arXiv2025

Enhancing Smart Contract Security Analysis with Execution Property Graphs

Smart contract vulnerabilities have led to significant financial losses, with their increasing complexity rendering outright prevention of hacks increasingly challenging. This trend highlights the crucial need for advanced forensic analysis and real-time intrusion detection, where dynamic analysis plays a key role in dissecting smart contract executions. Therefore, there is a pressing need for a unified and generic representation of smart contract executions, complemented by an efficient methodology that enables the modeling and identification of a broad spectrum of emerging attacks. We introduce Clue, a dynamic analysis framework specifically designed for the Ethereum virtual machine. Central to Clue is its ability to capture critical runtime information during contract executions, employing a novel graph-based representation, the Execution Property Graph. A key feature of Clue is its innovative graph traversal technique, which is adept at detecting complex attacks, including (read-only) reentrancy and price manipulation. Evaluation results reveal Clue's superior performance with high true positive rates and low false positive rates, outperforming state-of-the-art tools. Furthermore, Clue's efficiency positions it as a valuable tool for both forensic analysis and real-time intrusion detection.

cs.CR↗

arXiv2025

MMDT: Decoding the Trustworthiness and Safety of Multimodal Foundation Models

Multimodal foundation models (MMFMs) play a crucial role in various applications, including autonomous driving, healthcare, and virtual assistants. However, several studies have revealed vulnerabilities in these models, such as generating unsafe content by text-to-image models. Existing benchmarks on multimodal models either predominantly assess the helpfulness of these models, or only focus on limited perspectives such as fairness and privacy. In this paper, we present the first unified platform, MMDT (Multimodal DecodingTrust), designed to provide a comprehensive safety and trustworthiness evaluation for MMFMs. Our platform assesses models from multiple perspectives, including safety, hallucination, fairness/bias, privacy, adversarial robustness, and out-of-distribution (OOD) generalization. We have designed various evaluation scenarios and red teaming algorithms under different tasks for each perspective to generate challenging data, forming a high-quality benchmark. We evaluate a range of multimodal models using MMDT, and our findings reveal a series of vulnerabilities and areas for improvement across these perspectives. This work introduces the first comprehensive and unique safety and trustworthiness evaluation platform for MMFMs, paving the way for developing safer and more reliable MMFMs and systems. Our platform and benchmark are available at https://mmdecodingtrust.github.io/.

cs.CL↗

arXiv2025

Boosting Alignment for Post-Unlearning Text-to-Image Generative Models

Large-scale generative models have shown impressive image-generation capabilities, propelled by massive data. However, this often inadvertently leads to the generation of harmful or inappropriate content and raises copyright concerns. Driven by these concerns, machine unlearning has become crucial to effectively purge undesirable knowledge from models. While existing literature has studied various unlearning techniques, these often suffer from either poor unlearning quality or degradation in text-image alignment after unlearning, due to the competitive nature of these objectives. To address these challenges, we propose a framework that seeks an optimal model update at each unlearning iteration, ensuring monotonic improvement on both objectives. We further derive the characterization of such an update. In addition, we design procedures to strategically diversify the unlearning and remaining datasets to boost performance improvement. Our evaluation demonstrates that our method effectively removes target classes from recent diffusion-based generative models and concepts from stable diffusion models while maintaining close alignment with the models' original trained states, thus outperforming state-of-the-art baselines. Our code will be made available at https://github.com/reds-lab/Restricted_gradient_diversity_unlearning.git.

cs.LG↗

arXiv2024

LLM-PBE: Assessing Data Privacy in Large Language Models

Large Language Models (LLMs) have become integral to numerous domains, significantly advancing applications in data management, mining, and analysis. Their profound capabilities in processing and interpreting complex language data, however, bring to light pressing concerns regarding data privacy, especially the risk of unintentional training data leakage. Despite the critical nature of this issue, there has been no existing literature to offer a comprehensive assessment of data privacy risks in LLMs. Addressing this gap, our paper introduces LLM-PBE, a toolkit crafted specifically for the systematic evaluation of data privacy risks in LLMs. LLM-PBE is designed to analyze privacy across the entire lifecycle of LLMs, incorporating diverse attack and defense strategies, and handling various data types and metrics. Through detailed experimentation with multiple LLMs, LLM-PBE facilitates an in-depth exploration of data privacy concerns, shedding light on influential factors such as model size, data characteristics, and evolving temporal dimensions. This study not only enriches the understanding of privacy issues in LLMs but also serves as a vital resource for future research in the field. Aimed at enhancing the breadth of knowledge in this area, the findings, resources, and our full technical report are made available at https://llm-pbe.github.io/, providing an open platform for academic and practical advancements in LLM privacy assessment.

cs.CR↗

arXiv2023

Unmasking Role-Play Attack Strategies in Exploiting Decentralized Finance (DeFi) Systems

The rapid growth and adoption of decentralized finance (DeFi) systems have been accompanied by various threats, notably those emerging from vulnerabilities in their intricate design. In our work, we introduce and define an attack strategy termed as Role-Play Attack, in which the attacker acts as multiple roles concurrently to exploit the DeFi system and cause substantial financial losses. We provide a formal definition of this strategy and demonstrate its potential impacts by revealing the total loss of \$435.1M caused by 14 historical attacks with applying this pattern. Besides, we mathematically analyzed the attacks with top 2 losses and retrofitted the corresponding attack pattern by concrete execution, indicating that this strategy could increase the potential profit for original attacks by \$3.34M (51.4%) and \$3.76M (12.0%), respectively.

cs.CR↗

arXiv2020

Intelligent Credit Limit Management in Consumer Loans Based on Causal Inference

Nowadays consumer loan plays an important role in promoting the economic growth, and credit cards are the most popular consumer loan. One of the most essential parts in credit cards is the credit limit management. Traditionally, credit limits are adjusted based on limited heuristic strategies, which are developed by experienced professionals. In this paper, we present a data-driven approach to manage the credit limit intelligently. Firstly, a conditional independence testing is conducted to acquire the data for building models. Based on these testing data, a response model is then built to measure the heterogeneous treatment effect of increasing credit limits (i.e. treatments) for different customers, who are depicted by several control variables (i.e. features). In order to incorporate the diminishing marginal effect, a carefully selected log transformation is introduced to the treatment variable. Moreover, the model's capability can be further enhanced by applying a non-linear transformation on features via GBDT encoding. Finally, a well-designed metric is proposed to properly measure the performances of compared methods. The experimental results demonstrate the effectiveness of the proposed approach.

cs.LG↗