Search

Large Language Models (LLMs) have demonstrated impressive capabilities in reasoning using Chain-of-Thought (CoT) prompting. However, CoT can be biased by users' instruction. In this work, we study the reasoning robustness of LLMs to typographical errors, which can naturally occur in users' queries. We design an Adversarial Typo Attack ($\texttt{ATA}$) algorithm that iteratively samples typos for words that are important to the query and selects the edit that is most likely to succeed in attacking. It shows that LLMs are sensitive to minimal adversarial typographical changes. Notably, with 1 character edit, Mistral-7B-Instruct's accuracy drops from 43.7% to 38.6% on GSM8K, while with 8 character edits the performance further drops to 19.2%. To extend our evaluation to larger and closed-source LLMs, we develop the $\texttt{R$^2$ATA}$ benchmark, which assesses models' $\underline{R}$easoning $\underline{R}$obustness to $\underline{\texttt{ATA}}$. It includes adversarial typographical questions derived from three widely used reasoning datasets-GSM8K, BBH, and MMLU-by applying $\texttt{ATA}$ to open-source LLMs. $\texttt{R$^2$ATA}$ demonstrates remarkable transferability and causes notable performance drops across multiple super large and closed-source LLMs.

GPT-4o is an autoregressive omni model that accepts as input any combination of text, audio, image, and video, and generates any combination of text, audio, and image outputs. It's trained end-to-end across text, vision, and audio, meaning all inputs and outputs are processed by the same neural network. GPT-4o can respond to audio inputs in as little as 232 milliseconds, with an average of 320 milliseconds, which is similar to human response time in conversation. It matches GPT-4 Turbo performance on text in English and code, with significant improvement on text in non-English languages, while also being much faster and 50\% cheaper in the API. GPT-4o is especially better at vision and audio understanding compared to existing models. In line with our commitment to building AI safely and consistent with our voluntary commitments to the White House, we are sharing the GPT-4o System Card, which includes our Preparedness Framework evaluations. In this System Card, we provide a detailed look at GPT-4o's capabilities, limitations, and safety evaluations across multiple categories, focusing on speech-to-speech while also evaluating text and image capabilities, and measures we've implemented to ensure the model is safe and aligned. We also include third-party assessments on dangerous capabilities, as well as discussion of potential societal impacts of GPT-4o's text and vision capabilities.

Perceiving and generating diverse modalities are crucial for AI models to effectively learn from and engage with real-world signals, necessitating reliable evaluations for their development. We identify two major issues in current evaluations: (1) inconsistent standards, shaped by different communities with varying protocols and maturity levels; and (2) significant query, grading, and generalization biases. To address these, we introduce MixEval-X, the first any-to-any, real-world benchmark designed to optimize and standardize evaluations across diverse input and output modalities. We propose multi-modal benchmark mixture and adaptation-rectification pipelines to reconstruct real-world task distributions, ensuring evaluations generalize effectively to real-world use cases. Extensive meta-evaluations show our approach effectively aligns benchmark samples with real-world task distributions. Meanwhile, MixEval-X's model rankings correlate strongly with that of crowd-sourced real-world evaluations (up to 0.98) while being much more efficient. We provide comprehensive leaderboards to rerank existing models and organizations and offer insights to enhance understanding of multi-modal evaluations and inform future research.

Language Language Models (LLMs) face safety concerns due to potential misuse by malicious users. Recent red-teaming efforts have identified adversarial suffixes capable of jailbreaking LLMs using the gradient-based search algorithm Greedy Coordinate Gradient (GCG). However, GCG struggles with computational inefficiency, limiting further investigations regarding suffix transferability and scalability across models and data. In this work, we bridge the connection between search efficiency and suffix transferability. We propose a two-stage transfer learning framework, DeGCG, which decouples the search process into behavior-agnostic pre-searching and behavior-relevant post-searching. Specifically, we employ direct first target token optimization in pre-searching to facilitate the search process. We apply our approach to cross-model, cross-data, and self-transfer scenarios. Furthermore, we introduce an interleaved variant of our approach, i-DeGCG, which iteratively leverages self-transferability to accelerate the search process. Experiments on HarmBench demonstrate the efficiency of our approach across various models and domains. Notably, our i-DeGCG outperforms the baseline on Llama2-chat-7b with ASRs of $43.9$ ($+22.2$) and $39.0$ ($+19.5$) on valid and test sets, respectively. Further analysis on cross-model transfer indicates the pivotal role of first target token optimization in leveraging suffix transferability for efficient searching., Comment: Accepted to the EMNLP 2024

Large Language Models (LLMs) excel in stand-alone code tasks like HumanEval and MBPP, but struggle with handling entire code repositories. This challenge has prompted research on enhancing LLM-codebase interaction at a repository scale. Current solutions rely on similarity-based retrieval or manual tools and APIs, each with notable drawbacks. Similarity-based retrieval often has low recall in complex tasks, while manual tools and APIs are typically task-specific and require expert knowledge, reducing their generalizability across diverse code tasks and real-world applications. To mitigate these limitations, we introduce CodexGraph, a system that integrates LLM agents with graph database interfaces extracted from code repositories. By leveraging the structural properties of graph databases and the flexibility of the graph query language, CodexGraph enables the LLM agent to construct and execute queries, allowing for precise, code structure-aware context retrieval and code navigation. We assess CodexGraph using three benchmarks: CrossCodeEval, SWE-bench, and EvoCodeBench. Additionally, we develop five real-world coding applications. With a unified graph database schema, CodexGraph demonstrates competitive performance and potential in both academic and real-world environments, showcasing its versatility and efficacy in software engineering. Our application demo: https://github.com/modelscope/modelscope-agent/tree/master/apps/codexgraph_agent., Comment: work in progress

Antiferromagnetic intermetallic compound EuRhGe3 crystalizes in a non-centrosymmetric BaNiSn3-type (I4mm) structure. We studied its pressure-dependent crystal structure by using synchrotron powder x-ray diffraction at room temperature. Our results show a smooth contraction of the unit cell volume by applying pressure while preserving I4mm symmetry. No structural transition was observed up to 35 GPa. By the equation of state fitting analysis, the bulk modulus and its pressure derivative were determined to be 73 (1) GPa and 5.5 (2), respectively. Furthermore, similar to the isostructural EuCoGe3, an anisotropic compression of a and c lattice parameters was observed. Our experimental results show a good agreement with the pressure-dependent structural evolution expected from theoretical calculations below 13 GPa. Reflecting a strong deviation from integer Eu valence, the experimental volume data appear to be smaller than those of DFT calculated values at higher pressures.

This research was devoted to investigate the inverse spectral problem of Sturm-Liouville operator with many frozen arguments. Under some assumptions, the authors obtained uniqueness theorems. At the end, a numerical simulation for the inverse problem was presented., Comment: 6 pages with 2 figures

When LLMs are deployed in sensitive, human-facing settings, it is crucial that they do not output unsafe, biased, or privacy-violating outputs. For this reason, models are both trained and instructed to refuse to answer unsafe prompts such as "Tell me how to build a bomb." We find that, despite these safeguards, it is possible to break model defenses simply by appending a space to the end of a model's input. In a study of eight open-source models, we demonstrate that this acts as a strong enough attack to cause the majority of models to generate harmful outputs with very high success rates. We examine the causes of this behavior, finding that the contexts in which single spaces occur in tokenized training data encourage models to generate lists when prompted, overriding training signals to refuse to answer unsafe requests. Our findings underscore the fragile state of current model alignment and promote the importance of developing more robust alignment methods. Code and data will be available at https://github.com/hannah-aught/space_attack., Comment: 8 pages, 6 figures

We introduce a defense against adversarial attacks on LLMs utilizing self-evaluation. Our method requires no model fine-tuning, instead using pre-trained models to evaluate the inputs and outputs of a generator model, significantly reducing the cost of implementation in comparison to other, finetuning-based methods. Our method can significantly reduce the attack success rate of attacks on both open and closed-source LLMs, beyond the reductions demonstrated by Llama-Guard2 and commonly used content moderation APIs. We present an analysis of the effectiveness of our method, including attempts to attack the evaluator in various settings, demonstrating that it is also more resilient to attacks than existing methods. Code and data will be made available at https://github.com/Linlt-leon/self-eval., Comment: 8 pages, 7 figures

Extrapolation in Large language models (LLMs) for open-ended inquiry encounters two pivotal issues: (1) hallucination and (2) expensive training costs. These issues present challenges for LLMs in specialized domains and personalized data, requiring truthful responses and low fine-tuning costs. Existing works attempt to tackle the problem by augmenting the input of a smaller language model with information from a knowledge graph (KG). However, they have two limitations: (1) failing to extract relevant information from a large one-hop neighborhood in KG and (2) applying the same augmentation strategy for KGs with different characteristics that may result in low performance. Moreover, open-ended inquiry typically yields multiple responses, further complicating extrapolation. We propose a new task, the extreme multi-label KG link prediction task, to enable a model to perform extrapolation with multiple responses using structured real-world knowledge. Our retriever identifies relevant one-hop neighbors by considering entity, relation, and textual data together. Our experiments demonstrate that (1) KGs with different characteristics require different augmenting strategies, and (2) augmenting the language model's input with textual data improves task performance significantly. By incorporating the retrieval-augmented framework with KG, our framework, with a small parameter size, is able to extrapolate based on a given KG. The code can be obtained on GitHub: https://github.com/exiled1143/Retrieval-Augmented-Language-Model-for-Multi-Label-Knowledge-Graph-Link-Prediction.git

This project addresses the challenge of human motion prediction, a critical area for applications such as au- tonomous vehicle movement detection. Previous works have emphasized the need for low inference times to provide real time performance for applications like these. Our primary objective is to critically evaluate existing model ar- chitectures, identifying their advantages and opportunities for improvement by replicating the state-of-the-art (SOTA) Spatio-Temporal Transformer model as best as possible given computational con- straints. These models have surpassed the limitations of RNN-based models and have demonstrated the ability to generate plausible motion sequences over both short and long term horizons through the use of spatio-temporal rep- resentations. We also propose a novel architecture to ad- dress challenges of real time inference speed by incorpo- rating a Mixture of Experts (MoE) block within the Spatial- Temporal (ST) attention layer. The particular variation that is used is Soft MoE, a fully-differentiable sparse Transformer that has shown promising ability to enable larger model capacity at lower inference cost. We make out code publicly available at https://github.com/edshieh/motionprediction, Comment: 16 pages, 6 figures

The rapid emergence of generative Language Models (LMs) has led to growing concern about the impacts that their unexamined adoption may have on the social well-being of diverse user groups. Meanwhile, LMs are increasingly being adopted in K-20 schools and one-on-one student settings with minimal investigation of potential harms associated with their deployment. Motivated in part by real-world/everyday use cases (e.g., an AI writing assistant) this paper explores the potential psychosocial harms of stories generated by five leading LMs in response to open-ended prompting. We extend findings of stereotyping harms analyzing a total of 150K 100-word stories related to student classroom interactions. Examining patterns in LM-generated character demographics and representational harms (i.e., erasure, subordination, and stereotyping) we highlight particularly egregious vignettes, illustrating the ways LM-generated outputs may influence the experiences of users with marginalized and minoritized identities, and emphasizing the need for a critical understanding of the psychosocial impacts of generative AI tools when deployed and utilized in diverse social contexts., Comment: Presented in Impact of GenAI on Social and Individual Well-being at AAAI 2024 Spring Symposium Series (2024)

We introduce an approach aimed at enhancing the reasoning capabilities of Large Language Models (LLMs) through an iterative preference learning process inspired by the successful strategy employed by AlphaZero. Our work leverages Monte Carlo Tree Search (MCTS) to iteratively collect preference data, utilizing its look-ahead ability to break down instance-level rewards into more granular step-level signals. To enhance consistency in intermediate steps, we combine outcome validation and stepwise self-evaluation, continually updating the quality assessment of newly generated data. The proposed algorithm employs Direct Preference Optimization (DPO) to update the LLM policy using this newly generated step-level preference data. Theoretical analysis reveals the importance of using on-policy sampled data for successful self-improving. Extensive evaluations on various arithmetic and commonsense reasoning tasks demonstrate remarkable performance improvements over existing models. For instance, our approach outperforms the Mistral-7B Supervised Fine-Tuning (SFT) baseline on GSM8K, MATH, and ARC-C, with substantial increases in accuracy to $81.8\%$ (+$5.9\%$), $34.7\%$ (+$5.8\%$), and $76.4\%$ (+$15.8\%$), respectively. Additionally, our research delves into the training and inference compute tradeoff, providing insights into how our method effectively maximizes performance gains. Our code is publicly available at https://github.com/YuxiXie/MCTS-DPO., Comment: 10 pages, 4 figures, 4 tables (24 pages, 9 figures, 9 tables including references and appendices)

M\"ossbauer spectroscopy is a technique employed to investigate the microscopic properties of materials using transitions between energy levels in the nuclei. Conventionally, in synchrotron-radiation-based M\"ossbauer spectroscopy, the measurement window is decided by the researcher heuristically, although this decision has a significant impact on the shape of the measurement spectra. In this paper, we propose a method for evaluating the precision of the spectral position by introducing Bayesian estimation. The proposed method makes it possible to select the best measurement window by calculating the precision of M\"ossbauer spectroscopy from the data. Based on the results, the precision of the M\"ossbauer center shifts improved by more than three times compared with the results achieved with the conventional simple fitting method using the Lorentzian function., Comment: 8 pages, 8 figures

The rapid deployment of generative language models (LMs) has raised concerns about social biases affecting the well-being of diverse consumers. The extant literature on generative LMs has primarily examined bias via explicit identity prompting. However, prior research on bias in earlier language-based technology platforms, including search engines, has shown that discrimination can occur even when identity terms are not specified explicitly. Studies of bias in LM responses to open-ended prompts (where identity classifications are left unspecified) are lacking and have not yet been grounded in end-consumer harms. Here, we advance studies of generative LM bias by considering a broader set of natural use cases via open-ended prompting. In this "laissez-faire" setting, we find that synthetically generated texts from five of the most pervasive LMs (ChatGPT3.5, ChatGPT4, Claude2.0, Llama2, and PaLM2) perpetuate harms of omission, subordination, and stereotyping for minoritized individuals with intersectional race, gender, and/or sexual orientation identities (AI/AN, Asian, Black, Latine, MENA, NH/PI, Female, Non-binary, Queer). We find widespread evidence of bias to an extent that such individuals are hundreds to thousands of times more likely to encounter LM-generated outputs that portray their identities in a subordinated manner compared to representative or empowering portrayals. We also document a prevalence of stereotypes (e.g. perpetual foreigner) in LM-generated outputs that are known to trigger psychological harms that disproportionately affect minoritized individuals. These include stereotype threat, which leads to impaired cognitive performance and increased negative self-perception. Our findings highlight the urgent need to protect consumers from discriminatory harms caused by language models and invest in critical AI education programs tailored towards empowering diverse consumers., Comment: 16 pages (43 if including supplementals), 8 figures (23 if including supplementals)

This paper introduces a novel approach to optimize energy efficiency in wavelength division multiplexing (WDM) Visible Light Communication (VLC) systems designed for indoor broadcasting networks. A physics-based LED model is integrated into system energy efficiency optimization, enabling quantitative analysis of the critical issue of VLC energy efficiency: the nonlinear interplay between illumination and communication performance. The optimization jointly incorporates constraints on communication quality of each channel, and illumination performance, standardized by the International Commission on Illumination (CIE). The formulated nonlinear optimization problem is solved by the Sequential Quadratic Programming (SQP) algorithm in an experiment-based simulation. An integrated Red-Green-Blue-Yellow Light Emitting Diode (RGBY-LED) is measured for model calibration and three different scenarios are simulated to evaluate the generality of the proposed method. Results demonstrate a double enhancement in performance and a high versatility in accommodating various scenarios. Furthermore, it highlights the importance of balancing communication and illumination imperatives in VLC systems, challenging conventional perceptions focused solely on minimizing power consumption.

Transformer-based models have gained widespread popularity in both the computer vision (CV) and natural language processing (NLP) fields. However, significant challenges arise during post-training linear quantization, leading to noticeable reductions in inference accuracy. Our study focuses on uncovering the underlying causes of these accuracy drops and proposing a quantization-friendly fine-tuning method, \textbf{QuantTune}. Firstly, our analysis revealed that, on average, 65\% of quantization errors result from the precision loss incurred by the dynamic range amplification effect of outliers across the target Transformer-based models. Secondly, \textbf{QuantTune} adjusts weights based on the deviation of outlier activations and effectively constrains the dynamic ranges of the problematic activations. As a result, it successfully mitigates the negative impact of outliers on the inference accuracy of quantized models. Lastly, \textbf{QuantTune} can be seamlessly integrated into the back-propagation pass in the fine-tuning process without requiring extra complexity in inference software and hardware design. Our approach showcases significant improvements in post-training quantization across a range of Transformer-based models, including ViT, Bert-base, and OPT. QuantTune reduces accuracy drops by 12.09\% at 8-bit quantization and 33.8\% at 7-bit compared to top calibration methods, outperforming state-of-the-art solutions by over 18.84\% across ViT models.

Safety of Large Language Models (LLMs) has become a critical issue given their rapid progresses. Greedy Coordinate Gradient (GCG) is shown to be effective in constructing adversarial prompts to break the aligned LLMs, but optimization of GCG is time-consuming. To reduce the time cost of GCG and enable more comprehensive studies of LLM safety, in this work, we study a new algorithm called $\texttt{Probe sampling}$. At the core of the algorithm is a mechanism that dynamically determines how similar a smaller draft model's predictions are to the target model's predictions for prompt candidates. When the target model is similar to the draft model, we rely heavily on the draft model to filter out a large number of potential prompt candidates. Probe sampling achieves up to $5.6$ times speedup using Llama2-7b-chat and leads to equal or improved attack success rate (ASR) on the AdvBench. Furthermore, probe sampling is also able to accelerate other prompt optimization techniques and adversarial methods, leading to acceleration of $1.8\times$ for AutoPrompt, $2.4\times$ for APE and $2.4\times$ for AutoDAN.

This paper examines the application of WiFi signals for real-world monitoring of daily activities in home healthcare scenarios. While the state-of-the-art of WiFi-based activity recognition is promising in lab environments, challenges arise in real-world settings due to environmental, subject, and system configuration variables, affecting accuracy and adaptability. The research involved deploying systems in various settings and analyzing data shifts. It aims to guide realistic development of robust, context-aware WiFi sensing systems for elderly care. The findings suggest a shift in WiFi-based activity sensing, bridging the gap between academic research and practical applications, enhancing life quality through technology.

The skeletal region is one of the common sites of metastatic spread of cancer in the breast and prostate. CT is routinely used to measure the size of lesions in the bones. However, they can be difficult to spot due to the wide variations in their sizes, shapes, and appearances. Precise localization of such lesions would enable reliable tracking of interval changes (growth, shrinkage, or unchanged status). To that end, an automated technique to detect bone lesions is highly desirable. In this pilot work, we developed a pipeline to detect bone lesions (lytic, blastic, and mixed) in CT volumes via a proxy segmentation task. First, we used the bone lesions that were prospectively marked by radiologists in a few 2D slices of CT volumes and converted them into weak 3D segmentation masks. Then, we trained a 3D full-resolution nnUNet model using these weak 3D annotations to segment the lesions and thereby detected them. Our automated method detected bone lesions in CT with a precision of 96.7% and recall of 47.3% despite the use of incomplete and partial training data. To the best of our knowledge, we are the first to attempt the direct detection of bone lesions in CT via a proxy segmentation task., Comment: Accepted at SPIE 2024

We propose a new method, Adversarial In-Context Learning (adv-ICL), to optimize prompt for in-context learning (ICL) by employing one LLM as a generator, another as a discriminator, and a third as a prompt modifier. As in traditional adversarial learning, adv-ICL is implemented as a two-player game between the generator and discriminator, where the generator tries to generate realistic enough output to fool the discriminator. In each round, given an input prefixed by task instructions and several exemplars, the generator produces an output. The discriminator is then tasked with classifying the generator input-output pair as model-generated or real data. Based on the discriminator loss, the prompt modifier proposes possible edits to the generator and discriminator prompts, and the edits that most improve the adversarial loss are selected. We show that adv-ICL results in significant improvements over state-of-the-art prompt optimization techniques for both open and closed-source models on 11 generation and classification tasks including summarization, arithmetic reasoning, machine translation, data-to-text generation, and the MMLU and big-bench hard benchmarks. In addition, because our method uses pre-trained models and updates only prompts rather than model parameters, it is computationally efficient, easy to extend to any LLM and task, and effective in low-resource settings., Comment: ACL 2024

Fr\'echet Inception Distance (FID) is a widely used metric for assessing synthetic image quality. It relies on an ImageNet-based feature extractor, making its applicability to medical imaging unclear. A recent trend is to adapt FID to medical imaging through feature extractors trained on medical images. Our study challenges this practice by demonstrating that ImageNet-based extractors are more consistent and aligned with human judgment than their RadImageNet counterparts. We evaluated sixteen StyleGAN2 networks across four medical imaging modalities and four data augmentation techniques with Fr\'echet distances (FDs) computed using eleven ImageNet or RadImageNet-trained feature extractors. Comparison with human judgment via visual Turing tests revealed that ImageNet-based extractors produced rankings consistent with human judgment, with the FD derived from the ImageNet-trained SwAV extractor significantly correlating with expert evaluations. In contrast, RadImageNet-based rankings were volatile and inconsistent with human judgment. Our findings challenge prevailing assumptions, providing novel evidence that medical image-trained feature extractors do not inherently improve FDs and can even compromise their reliability. Our code is available at https://github.com/mckellwoodland/fid-med-eval., Comment: This preprint has not undergone peer review or any post-submission improvements or corrections. The Version of Record of this contribution is published in LNCS vol. 15012, and is available online at https://doi.org/10.1007/978-3-031-72390-2_9

The NA48/2 experiment at CERN reports the first observation of the $K^{\pm} \rightarrow \pi^{0} \pi^{0} \mu^{\pm} \nu$ decay based on a sample of 2437 candidates with 15% background contamination collected in 2003--2004. The decay branching ratio in the kinematic region of the squared dilepton mass above $0.03$~GeV$^2/c^4$ is measured to be $(0.65 \pm 0.03) \times 10^{-6}$. The extrapolation to the full kinematic space, using a specific model, is found to be $(3.45 \pm 0.16) \times 10^{-6}$, in agreement with chiral perturbation theory predictions., Comment: 16 pages, 9 figures

Accurately measuring the evolution of Multiple Sclerosis (MS) with magnetic resonance imaging (MRI) critically informs understanding of disease progression and helps to direct therapeutic strategy. Deep learning models have shown promise for automatically segmenting MS lesions, but the scarcity of accurately annotated data hinders progress in this area. Obtaining sufficient data from a single clinical site is challenging and does not address the heterogeneous need for model robustness. Conversely, the collection of data from multiple sites introduces data privacy concerns and potential label noise due to varying annotation standards. To address this dilemma, we explore the use of the federated learning framework while considering label noise. Our approach enables collaboration among multiple clinical sites without compromising data privacy under a federated learning paradigm that incorporates a noise-robust training strategy based on label correction. Specifically, we introduce a Decoupled Hard Label Correction (DHLC) strategy that considers the imbalanced distribution and fuzzy boundaries of MS lesions, enabling the correction of false annotations based on prediction confidence. We also introduce a Centrally Enhanced Label Correction (CELC) strategy, which leverages the aggregated central model as a correction teacher for all sites, enhancing the reliability of the correction process. Extensive experiments conducted on two multi-site datasets demonstrate the effectiveness and robustness of our proposed methods, indicating their potential for clinical applications in multi-site collaborations., Comment: 11 pages, 4 figures, journal submission