Abstract:Large Language Models (LLMs) are employed across various high-stakes domains, where the reliability of their outputs is crucial. One commonly used method to assess the reliability of LLMs' responses is uncertainty estimation, which gauges the likelihood of their answers being correct. While many studies focus on improving the accuracy of uncertainty estimations for LLMs, our research investigates the fragility of uncertainty estimation and explores potential attacks. We demonstrate that an attacker can embed a backdoor in LLMs, which, when activated by a specific trigger in the input, manipulates the model's uncertainty without affecting the final output. Specifically, the proposed backdoor attack method can alter an LLM's output probability distribution, causing the probability distribution to converge towards an attacker-predefined distribution while ensuring that the top-1 prediction remains unchanged. Our experimental results demonstrate that this attack effectively undermines the model's self-evaluation reliability in multiple-choice questions. For instance, we achieved a 100 attack success rate (ASR) across three different triggering strategies in four models. Further, we investigate whether this manipulation generalizes across different prompts and domains. This work highlights a significant threat to the reliability of LLMs and underscores the need for future defenses against such attacks. The code is available at https://github.com/qcznlp/uncertainty_attack.
Abstract:In the rapidly evolving field of artificial intelligence, large language models (LLMs) have emerged as powerful tools for a myriad of applications, from natural language processing to decision-making support systems. However, as these models become increasingly integrated into societal frameworks, the imperative to ensure they operate within ethical and moral boundaries has never been more critical. This paper introduces a novel benchmark designed to measure and compare the moral reasoning capabilities of LLMs. We present the first comprehensive dataset specifically curated to probe the moral dimensions of LLM outputs, addressing a wide range of ethical dilemmas and scenarios reflective of real-world complexities. The main contribution of this work lies in the development of benchmark datasets and metrics for assessing the moral identity of LLMs, which accounts for nuance, contextual sensitivity, and alignment with human ethical standards. Our methodology involves a multi-faceted approach, combining quantitative analysis with qualitative insights from ethics scholars to ensure a thorough evaluation of model performance. By applying our benchmark across several leading LLMs, we uncover significant variations in moral reasoning capabilities of different models. These findings highlight the importance of considering moral reasoning in the development and evaluation of LLMs, as well as the need for ongoing research to address the biases and limitations uncovered in our study. We publicly release the benchmark at https://drive.google.com/drive/u/0/folders/1k93YZJserYc2CkqP8d4B3M3sgd3kA8W7 and also open-source the code of the project at https://github.com/agiresearch/MoralBench.
Abstract:This study intends to systematically disentangle pure logic reasoning and text understanding by investigating the contrast across abstract and contextualized logical problems from a comprehensive set of domains. We explore whether LLMs demonstrate genuine reasoning capabilities across various domains when the underlying logical structure remains constant. We focus on two main questions (1) Can abstract logical problems alone accurately benchmark an LLM's reasoning ability in real-world scenarios, disentangled from contextual support in practical settings? (2) Does fine-tuning LLMs on abstract logic problem generalize to contextualized logic problems and vice versa? To investigate these questions, we focus on standard propositional logic, specifically propositional deductive and abductive logic reasoning. In particular, we construct instantiated datasets for deductive and abductive reasoning with 4 levels of difficulty, encompassing 12 distinct categories or domains based on the categorization of Wikipedia. Our experiments aim to provide insights into disentangling context in logical reasoning and the true reasoning capabilities of LLMs and their generalization potential. The code and dataset are available at: https://github.com/agiresearch/ContextHub.
Abstract:Self-supervised learning models are vulnerable to backdoor attacks. Existing backdoor attacks that are effective in self-supervised learning often involve noticeable triggers, like colored patches, which are vulnerable to human inspection. In this paper, we propose an imperceptible and effective backdoor attack against self-supervised models. We first find that existing imperceptible triggers designed for supervised learning are not as effective in compromising self-supervised models. We then identify this ineffectiveness is attributed to the overlap in distributions between the backdoor and augmented samples used in self-supervised learning. Building on this insight, we design an attack using optimized triggers that are disentangled to the augmented transformation in the self-supervised learning, while also remaining imperceptible to human vision. Experiments on five datasets and seven SSL algorithms demonstrate our attack is highly effective and stealthy. It also has strong resistance to existing backdoor defenses. Our code can be found at https://github.com/Zhang-Henry/IMPERATIVE.
Abstract:This paper presents BattleAgent, an emulation system that combines the Large Vision-Language Model and Multi-agent System. This novel system aims to simulate complex dynamic interactions among multiple agents, as well as between agents and their environments, over a period of time. It emulates both the decision-making processes of leaders and the viewpoints of ordinary participants, such as soldiers. The emulation showcases the current capabilities of agents, featuring fine-grained multi-modal interactions between agents and landscapes. It develops customizable agent structures to meet specific situational requirements, for example, a variety of battle-related activities like scouting and trench digging. These components collaborate to recreate historical events in a lively and comprehensive manner while offering insights into the thoughts and feelings of individuals from diverse viewpoints. The technological foundations of BattleAgent establish detailed and immersive settings for historical battles, enabling individual agents to partake in, observe, and dynamically respond to evolving battle scenarios. This methodology holds the potential to substantially deepen our understanding of historical events, particularly through individual accounts. Such initiatives can also aid historical research, as conventional historical narratives often lack documentation and prioritize the perspectives of decision-makers, thereby overlooking the experiences of ordinary individuals. BattelAgent illustrates AI's potential to revitalize the human aspect in crucial social events, thereby fostering a more nuanced collective understanding and driving the progressive development of human society.
Abstract:This paper studies the phenomenon that different concepts are learned in different layers of large language models, i.e. more difficult concepts are fully acquired with deeper layers. We define the difficulty of concepts by the level of abstraction, and here it is crudely categorized by factual, emotional, and inferential. Each category contains a spectrum of tasks, arranged from simple to complex. For example, within the factual dimension, tasks range from lie detection to categorizing mathematical problems. We employ a probing technique to extract representations from different layers of the model and apply these to classification tasks. Our findings reveal that models tend to efficiently classify simpler tasks, indicating that these concepts are learned in shallower layers. Conversely, more complex tasks may only be discernible at deeper layers, if at all. This paper explores the implications of these findings for our understanding of model learning processes and internal representations. Our implementation is available at \url{https://github.com/Luckfort/CD}.
Abstract:Current large language models (LLMs) provide a strong foundation for large-scale user-oriented natural language tasks. A large number of users can easily inject adversarial text or instructions through the user interface, thus causing LLMs model security challenges. Although there is currently a large amount of research on prompt injection attacks, most of these black-box attacks use heuristic strategies. It is unclear how these heuristic strategies relate to the success rate of attacks and thus effectively improve model robustness. To solve this problem, we redefine the goal of the attack: to maximize the KL divergence between the conditional probabilities of the clean text and the adversarial text. Furthermore, we prove that maximizing the KL divergence is equivalent to maximizing the Mahalanobis distance between the embedded representation $x$ and $x'$ of the clean text and the adversarial text when the conditional probability is a Gaussian distribution and gives a quantitative relationship on $x$ and $x'$. Then we designed a simple and effective goal-guided generative prompt injection strategy (G2PIA) to find an injection text that satisfies specific constraints to achieve the optimal attack effect approximately. It is particularly noteworthy that our attack method is a query-free black-box attack method with low computational cost. Experimental results on seven LLM models and four datasets show the effectiveness of our attack method.
Abstract:Large Language Models (LLMs) have rapidly become important tools in Biomedical and Health Informatics (BHI), enabling new ways to analyze data, treat patients, and conduct research. This bibliometric review aims to provide a panoramic view of how LLMs have been used in BHI by examining research articles and collaboration networks from 2022 to 2023. It further explores how LLMs can improve Natural Language Processing (NLP) applications in various BHI areas like medical diagnosis, patient engagement, electronic health record management, and personalized medicine. To do this, our bibliometric review identifies key trends, maps out research networks, and highlights major developments in this fast-moving field. Lastly, it discusses the ethical concerns and practical challenges of using LLMs in BHI, such as data privacy and reliable medical recommendations. Looking ahead, we consider how LLMs could further transform biomedical research as well as healthcare delivery and patient outcomes. This bibliometric review serves as a resource for stakeholders in healthcare, including researchers, clinicians, and policymakers, to understand the current state and future potential of LLMs in BHI.
Abstract:The task of predicting multiple links within knowledge graphs (KGs) stands as a challenge in the field of knowledge graph analysis, a challenge increasingly resolvable due to advancements in natural language processing (NLP) and KG embedding techniques. This paper introduces a novel methodology, the Knowledge Graph Large Language Model Framework (KG-LLM), which leverages pivotal NLP paradigms, including chain-of-thought (CoT) prompting and in-context learning (ICL), to enhance multi-hop link prediction in KGs. By converting the KG to a CoT prompt, our framework is designed to discern and learn the latent representations of entities and their interrelations. To show the efficacy of the KG-LLM Framework, we fine-tune three leading Large Language Models (LLMs) within this framework, employing both non-ICL and ICL tasks for a comprehensive evaluation. Further, we explore the framework's potential to provide LLMs with zero-shot capabilities for handling previously unseen prompts. Our experimental findings discover that integrating ICL and CoT not only augments the performance of our approach but also significantly boosts the models' generalization capacity, thereby ensuring more precise predictions in unfamiliar scenarios.
Abstract:Understanding the reasoning capabilities of Multimodal Large Language Models (MLLMs) is an important area of research. In this study, we introduce a dynamic benchmark, NPHardEval4V, aimed at addressing the existing gaps in evaluating the pure reasoning abilities of MLLMs. Our benchmark aims to provide a venue to disentangle the effect of various factors such as image recognition and instruction following, from the overall performance of the models, allowing us to focus solely on evaluating their reasoning abilities. It is built by converting textual description of questions from NPHardEval to image representations. Our findings reveal significant discrepancies in reasoning abilities across different models and highlight the relatively weak performance of MLLMs compared to LLMs in terms of reasoning. We also investigate the impact of different prompting styles, including visual, text, and combined visual and text prompts, on the reasoning abilities of MLLMs, demonstrating the different impacts of multimodal inputs in model performance. Unlike traditional benchmarks, which focus primarily on static evaluations, our benchmark will be updated monthly to prevent overfitting and ensure a more authentic and fine-grained evaluation of the models. We believe that this benchmark can aid in understanding and guide the further development of reasoning abilities in MLLMs. The benchmark dataset and code are available at https://github.com/lizhouf/NPHardEval4V