In our research, we pioneer a novel approach to evaluate the effectiveness of jailbreak attacks on Large Language Models (LLMs), such as GPT-4 and LLaMa2, diverging from traditional robustness-focused binary evaluations. Our study introduces two distinct evaluation frameworks: a coarse-grained evaluation and a fine-grained evaluation. Each framework, using a scoring range from 0 to 1, offers a unique perspective, enabling a more comprehensive and nuanced evaluation of attack effectiveness and empowering attackers to refine their attack prompts with greater understanding. Furthermore, we have developed a comprehensive ground truth dataset specifically tailored for jailbreak tasks. This dataset not only serves as a crucial benchmark for our current study but also establishes a foundational resource for future research, enabling consistent and comparative analyses in this evolving field. Upon meticulous comparison with traditional evaluation methods, we discovered that our evaluation aligns with the baseline's trend while offering a more profound and detailed assessment. We believe that by accurately evaluating the effectiveness of attack prompts in the Jailbreak task, our work lays a solid foundation for assessing a wider array of similar or even more complex tasks in the realm of prompt injection, potentially revolutionizing this field.
相關內容
Solution for Emotion Prediction Competition of Workshop on Emotionally and Culturally Intelligent AI
This report provide a detailed description of the method that we explored and proposed in the WECIA Emotion Prediction Competition (EPC), which predicts a person's emotion through an artistic work with a comment. The dataset of this competition is ArtELingo, designed to encourage work on diversity across languages and cultures. The dataset has two main challenges, namely modal imbalance problem and language-cultural differences problem. In order to address this issue, we propose a simple yet effective approach called single-multi modal with Emotion-Cultural specific prompt(ECSP), which focuses on using the single modal message to enhance the performance of multimodal models and a well-designed prompt to reduce cultural differences problem. To clarify, our approach contains two main blocks: (1)XLM-R\cite{conneau2019unsupervised} based unimodal model and X$^2$-VLM\cite{zeng2022x} based multimodal model (2) Emotion-Cultural specific prompt. Our approach ranked first in the final test with a score of 0.627.
As jurisdictions around the world take their first steps toward regulating the most powerful AI systems, such as the EU AI Act and the US Executive Order 14110, there is a growing need for effective enforcement mechanisms that can verify compliance and respond to violations. We argue that compute providers should have legal obligations and ethical responsibilities associated with AI development and deployment, both to provide secure infrastructure and to serve as intermediaries for AI regulation. Compute providers can play an essential role in a regulatory ecosystem via four key capacities: as securers, safeguarding AI systems and critical infrastructure; as record keepers, enhancing visibility for policymakers; as verifiers of customer activities, ensuring oversight; and as enforcers, taking actions against rule violations. We analyze the technical feasibility of performing these functions in a targeted and privacy-conscious manner and present a range of technical instruments. In particular, we describe how non-confidential information, to which compute providers largely already have access, can provide two key governance-relevant properties of a computational workload: its type-e.g., large-scale training or inference-and the amount of compute it has consumed. Using AI Executive Order 14110 as a case study, we outline how the US is beginning to implement record keeping requirements for compute providers. We also explore how verification and enforcement roles could be added to establish a comprehensive AI compute oversight scheme. We argue that internationalization will be key to effective implementation, and highlight the critical challenge of balancing confidentiality and privacy with risk mitigation as the role of compute providers in AI regulation expands.
This study delves into the potential use of Large Language Models (LLMs) for generating Library of Congress Subject Headings (LCSH). The authors employed ChatGPT to generate subject headings for electronic theses and dissertations (ETDs) based on their titles and summaries. The results revealed that although some generated subject headings were valid, there were issues regarding specificity and exhaustiveness. The study showcases that LLMs can serve as a strategic response to the backlog of items awaiting cataloging in academic libraries, while also offering a cost-effective approach for promptly generating LCSH. Nonetheless, human catalogers remain essential for verifying and enhancing the validity, exhaustiveness, and specificity of LCSH generated by LLMs.
The use of Large Language Models (LLMs) for writing has sparked controversy both among readers and writers. On one hand, writers are concerned that LLMs will deprive them of agency and ownership, and readers are concerned about spending their time on text generated by soulless machines. On the other hand, AI-assistance can improve writing as long as writers can conform to publisher policies, and as long as readers can be assured that a text has been verified by a human. We argue that a system that captures the provenance of interaction with an LLM can help writers retain their agency, conform to policies, and communicate their use of AI to publishers and readers transparently. Thus we propose HaLLMark, a tool for visualizing the writer's interaction with the LLM. We evaluated HaLLMark with 13 creative writers, and found that it helped them retain a sense of control and ownership of the text.
In this paper, we introduce a novel convex formulation that seamlessly integrates the Material Point Method (MPM) with articulated rigid body dynamics in frictional contact scenarios. We extend the linear corotational hyperelastic model into the realm of elastoplasticity and include an efficient return mapping algorithm. This approach is particularly effective for MPM simulations involving significant deformation and topology changes, while preserving the convexity of the optimization problem. Our method ensures global convergence, enabling the use of large simulation time steps without compromising robustness. We have validated our approach through rigorous testing and performance evaluations, highlighting its superior capabilities in managing complex simulations relevant to robotics. Compared to previous MPM based robotic simulators, our method significantly improves the stability of contact resolution -- a critical factor in robot manipulation tasks. We make our method available in the open-source robotics toolkit, Drake.
In this paper, we propose a novel parameter and computation efficient tuning method for Multi-modal Large Language Models (MLLMs), termed Efficient Attention Skipping (EAS). Concretely, we first reveal that multi-head attentions (MHAs), the main computational overhead of MLLMs, are often redundant to downstream tasks. Based on this observation, EAS evaluates the attention redundancy and skips the less important MHAs to speed up inference. Besides, we also propose a novel propagation-of-information adapter (PIA) to serve the attention skipping of EAS and keep parameter efficiency, which can be further re-parameterized into feed-forward networks (FFNs) for zero-extra latency. To validate EAS, we apply it to a recently proposed MLLM called LaVIN and a classic VL pre-trained model called METER, and conduct extensive experiments on a set of benchmarks. The experiments show that EAS not only retains high performance and parameter efficiency, but also greatly speeds up inference speed. For instance, LaVIN-EAS can obtain 89.98\% accuracy on ScineceQA while speeding up inference by 2.2 times to LaVIN
We present a visual computing framework for analyzing moral rhetoric on social media around controversial topics. Using Moral Foundation Theory, we propose a methodology for deconstructing and visualizing the \textit{when}, \textit{where}, and \textit{who} behind each of these moral dimensions as expressed in microblog data. We characterize the design of this framework, developed in collaboration with experts from language processing, communications, and causal inference. Our approach integrates microblog data with multiple sources of geospatial and temporal data, and leverages unsupervised machine learning (generalized additive models) to support collaborative hypothesis discovery and testing. We implement this approach in a system named MOTIV. We illustrate this approach on two problems, one related to Stay-at-home policies during the COVID-19 pandemic, and the other related to the Black Lives Matter movement. Through detailed case studies and discussions with collaborators, we identify several insights discovered regarding the different drivers of moral sentiment in social media. Our results indicate that this visual approach supports rapid, collaborative hypothesis testing, and can help give insights into the underlying moral values behind controversial political issues. Supplemental Material: //osf.io/ygkzn/?view_only=6310c0886938415391d977b8aae8b749
This paper presents a comprehensive and practical guide for practitioners and end-users working with Large Language Models (LLMs) in their downstream natural language processing (NLP) tasks. We provide discussions and insights into the usage of LLMs from the perspectives of models, data, and downstream tasks. Firstly, we offer an introduction and brief summary of current GPT- and BERT-style LLMs. Then, we discuss the influence of pre-training data, training data, and test data. Most importantly, we provide a detailed discussion about the use and non-use cases of large language models for various natural language processing tasks, such as knowledge-intensive tasks, traditional natural language understanding tasks, natural language generation tasks, emergent abilities, and considerations for specific tasks.We present various use cases and non-use cases to illustrate the practical applications and limitations of LLMs in real-world scenarios. We also try to understand the importance of data and the specific challenges associated with each NLP task. Furthermore, we explore the impact of spurious biases on LLMs and delve into other essential considerations, such as efficiency, cost, and latency, to ensure a comprehensive understanding of deploying LLMs in practice. This comprehensive guide aims to provide researchers and practitioners with valuable insights and best practices for working with LLMs, thereby enabling the successful implementation of these models in a wide range of NLP tasks. A curated list of practical guide resources of LLMs, regularly updated, can be found at \url{//github.com/Mooler0410/LLMsPracticalGuide}.
Graph Neural Networks (GNNs) have been studied from the lens of expressive power and generalization. However, their optimization properties are less well understood. We take the first step towards analyzing GNN training by studying the gradient dynamics of GNNs. First, we analyze linearized GNNs and prove that despite the non-convexity of training, convergence to a global minimum at a linear rate is guaranteed under mild assumptions that we validate on real-world graphs. Second, we study what may affect the GNNs' training speed. Our results show that the training of GNNs is implicitly accelerated by skip connections, more depth, and/or a good label distribution. Empirical results confirm that our theoretical results for linearized GNNs align with the training behavior of nonlinear GNNs. Our results provide the first theoretical support for the success of GNNs with skip connections in terms of optimization, and suggest that deep GNNs with skip connections would be promising in practice.
In this paper, we propose a novel Feature Decomposition and Reconstruction Learning (FDRL) method for effective facial expression recognition. We view the expression information as the combination of the shared information (expression similarities) across different expressions and the unique information (expression-specific variations) for each expression. More specifically, FDRL mainly consists of two crucial networks: a Feature Decomposition Network (FDN) and a Feature Reconstruction Network (FRN). In particular, FDN first decomposes the basic features extracted from a backbone network into a set of facial action-aware latent features to model expression similarities. Then, FRN captures the intra-feature and inter-feature relationships for latent features to characterize expression-specific variations, and reconstructs the expression feature. To this end, two modules including an intra-feature relation modeling module and an inter-feature relation modeling module are developed in FRN. Experimental results on both the in-the-lab databases (including CK+, MMI, and Oulu-CASIA) and the in-the-wild databases (including RAF-DB and SFEW) show that the proposed FDRL method consistently achieves higher recognition accuracy than several state-of-the-art methods. This clearly highlights the benefit of feature decomposition and reconstruction for classifying expressions.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191