Search

Large language models are susceptible to jailbreak attacks, which can result in the generation of harmful content. While prior defenses mitigate these risks by perturbing or inspecting inputs, they ignore competing objectives, the underlying cause of alignment failures. In this paper, we propose Alignment-Enhanced Decoding (AED), a novel defense that employs adaptive decoding to address the root causes of jailbreak issues. We first define the Competitive Index to quantify alignment failures and utilize feedback from self-evaluation to compute post-alignment logits. Then, AED adaptively combines AED and post-alignment logits with the original logits to obtain harmless and helpful distributions. Consequently, our method enhances safety alignment while maintaining helpfulness. We conduct experiments across five models and four common jailbreaks, with the results validating the effectiveness of our approach. Code is available at https://github.com/GIGABaozi/AED.git., Comment: 15 pages, 5 figures

In this paper, we present MooER, a LLM-based large-scale automatic speech recognition (ASR) / automatic speech translation (AST) model of Moore Threads. A 5000h pseudo labeled dataset containing open source and self collected speech data is used for training. We achieve performance comparable to other open source models trained with up to hundreds of thousands of hours of labeled speech data. Meanwhile, experiments conducted on Covost2 Zh2en testset suggest that our model outperforms other open source Speech LLMs. A BLEU score of 25.2 can be obtained. The main contributions of this paper are summarized as follows. First, this paper presents a training strategy for encoders and LLMs on speech related tasks (including ASR and AST) using a small size of pseudo labeled data without any extra manual annotation and selection. Second, we release our ASR and AST models and plan to open-source our training code and strategy in the near future. Moreover, a model trained on 8wh scale training data is planned to be released later on.

Remote sensing image change detection (CD) is essential for analyzing land surface changes over time, with a significant challenge being the differentiation of actual changes from complex scenes while filtering out pseudo-changes. A primary contributor to this challenge is the intra-class dynamic changes due to phenological characteristics in natural areas. To overcome this, we introduce the InPhea model, which integrates phenological features into a remote sensing image CD framework. The model features a detector with a differential attention module for improved feature representation of change information, coupled with high-resolution feature extraction and spatial pyramid blocks to enhance performance. Additionally, a constrainer with four constraint modules and a multi-stage contrastive learning approach is employed to aid in the model's understanding of phenological characteristics. Experiments on the HRSCD, SECD, and PSCD-Wuhan datasets reveal that InPhea outperforms other models, confirming its effectiveness in addressing phenological pseudo-changes and its overall model superiority.

A leading approach to algorithm design aims to minimize the number of operations in an algorithm's compilation. One intuitively expects that reducing the number of operations may decrease the chance of errors. This paradigm is particularly prevalent in quantum computing, where gates are hard to implement and noise rapidly decreases a quantum computer's potential to outperform classical computers. Here, we find that minimizing the number of operations in a quantum algorithm can be counterproductive, leading to a noise sensitivity that induces errors when running the algorithm in non-ideal conditions. To show this, we develop a framework to characterize the resilience of an algorithm to perturbative noises (including coherent errors, dephasing, and depolarizing noise). Some compilations of an algorithm can be resilient against certain noise sources while being unstable against other noises. We condense these results into a tradeoff relation between an algorithm's number of operations and its noise resilience. We also show how this framework can be leveraged to identify compilations of an algorithm that are better suited to withstand certain noises.

Energy dissipation in collisionless plasmas is one of the most outstanding open questions in plasma physics. Magnetic reconnection and turbulence are two phenomena that can produce the conditions for energy dissipation. These two phenomena are closely related to each other in a wide range of plasmas. Turbulent fluctuations can emerge in critical regions of reconnection events, and magnetic reconnection can occur as a product of the turbulent cascade. In this study, we perform 2D particle-in-cell simulations of a reconnecting Harris current sheet in the presence of turbulent fluctuations to explore the effect of turbulence on the reconnection process in collisionless non-relativistic pair-plasmas. We find that the presence of a turbulent field can affect the onset and evolution of magnetic reconnection. Moreover, we observe the existence of a scale dependent amplitude of magnetic field fluctuations above which these fluctuations are able to disrupt the growing of magnetic islands. These fluctuations provide thermal energy to the particles within the current sheet and preferential perpendicular thermal energy to the background population.

Many artificial intelligence models process input data of different lengths and resolutions, making the shape of the tensors dynamic. The performance of these models depends on the shape of the tensors, which makes it difficult to optimize the tensors before the model runs. There are two common solutions to this problem. The first is to add useless data to the input to match a pre-optimized tensor library. The second is to use small basic tensors to create a tensor that is closest in size to the input data and then tune it to minimize padding. However, this second solution can be time-consuming. This paper proposes a new technique for deep learning compilers called FTuner. Instead of using a large design space or training a cost model, we use an abstract computational unit called the uKernel to patch together small, various-sized tensors to match the shape of the input tensor. We determine the shape of the uKernel using an analytic hardware information model. Experiments show that the FTuner can achieve comparable operators and end-to-end performance to vendor libraries and achieves 3\% speedup on existing auto-tuner with the model-training compiler while reducing tuning time by two orders of magnitude., Comment: 14 pages, 16 figures, 6 tables

We demonstrate that a single-color ultrashort optical pulse propagating in air can emit THz radiation along any direction with respect to its propagation axis. The emission angle can be adjusted by the flying focus technique which determines the speed and direction of the ionization front. When the ionization front velocity becomes superluminal, the THz emission corresponds to classical Cherenkov radiation., Comment: 8 pages, 5 figures

Image captioning (IC) systems, such as Microsoft Azure Cognitive Service, translate image content into descriptive language but can generate inaccuracies leading to misinterpretations. Advanced testing techniques like MetaIC and ROME aim to address these issues but face significant challenges. These methods require intensive manual labor for detailed annotations and often produce unrealistic images, either by adding unrelated objects or failing to remove existing ones. Additionally, they generate limited test suites, with MetaIC restricted to inserting specific objects and ROME limited to a narrow range of variations. We introduce SPOLRE, a novel automated tool for semantic-preserving object layout reconstruction in IC system testing. SPOLRE leverages four transformation techniques to modify object layouts without altering the image's semantics. This automated approach eliminates the need for manual annotations and creates realistic, varied test suites. Our tests show that over 75% of survey respondents find SPOLRE-generated images more realistic than those from state-of-the-art methods. SPOLRE excels in identifying caption errors, detecting 31,544 incorrect captions across seven IC systems with an average precision of 91.62%, surpassing other methods which average 85.65% accuracy and identify 17,160 incorrect captions. Notably, SPOLRE identified 6,236 unique issues within Azure, demonstrating its effectiveness against one of the most advanced IC systems.

In this report, we present RT-DETRv2, an improved Real-Time DEtection TRansformer (RT-DETR). RT-DETRv2 builds upon the previous state-of-the-art real-time detector, RT-DETR, and opens up a set of bag-of-freebies for flexibility and practicality, as well as optimizing the training strategy to achieve enhanced performance. To improve the flexibility, we suggest setting a distinct number of sampling points for features at different scales in the deformable attention to achieve selective multi-scale feature extraction by the decoder. To enhance practicality, we propose an optional discrete sampling operator to replace the grid_sample operator that is specific to RT-DETR compared to YOLOs. This removes the deployment constraints typically associated with DETRs. For the training strategy, we propose dynamic data augmentation and scale-adaptive hyperparameters customization to improve performance without loss of speed. Source code and pre-trained models will be available at https://github.com/lyuwenyu/RT-DETR.

Critical parameters are the key to superconductivity research, and reliable instrumentations can facilitate the study. Traditionally, one has to use several different measurement techniques to measure critical parameters separately. In this work, we develop the use of a single species of quantum sensor to determine and estimate several critical parameters with the help of independent simulation data. We utilize the nitrogen-vacancy (NV) center in the diamond, which recently emerged as a promising candidate for probing exotic features in condensed matter physics. The non-invasive and highly stable nature provides extraordinary opportunities to solve scientific problems in various systems. Using a high-quality single-crystalline YBa$_{2}$Cu$_{4}$O$_{8}$ (YBCO) as a platform, we demonstrate the use of diamond particles and a bulk diamond to probe the Meissner effect. The evolution of the vector magnetic field, the $H-T$ phase diagram, and the map of fluorescence contour are studied via NV sensing. Our results reveal different critical parameters, including lower critical field $H_{c1}$, upper critical field $H_{c2}$, and critical current density $j_{c}$, as well as verifying the unconventional nature of this high-temperature superconductor YBCO. Therefore, NV-based quantum sensing techniques have huge potential in condensed matter research.

Large Vision Language Models (VLMs) extend and enhance the perceptual abilities of Large Language Models (LLMs). Despite offering new possibilities for LLM applications, these advancements raise significant security and ethical concerns, particularly regarding the generation of harmful content. While LLMs have undergone extensive security evaluations with the aid of red teaming frameworks, VLMs currently lack a well-developed one. To fill this gap, we introduce Arondight, a standardized red team framework tailored specifically for VLMs. Arondight is dedicated to resolving issues related to the absence of visual modality and inadequate diversity encountered when transitioning existing red teaming methodologies from LLMs to VLMs. Our framework features an automated multi-modal jailbreak attack, wherein visual jailbreak prompts are produced by a red team VLM, and textual prompts are generated by a red team LLM guided by a reinforcement learning agent. To enhance the comprehensiveness of VLM security evaluation, we integrate entropy bonuses and novelty reward metrics. These elements incentivize the RL agent to guide the red team LLM in creating a wider array of diverse and previously unseen test cases. Our evaluation of ten cutting-edge VLMs exposes significant security vulnerabilities, particularly in generating toxic images and aligning multi-modal prompts. In particular, our Arondight achieves an average attack success rate of 84.5\% on GPT-4 in all fourteen prohibited scenarios defined by OpenAI in terms of generating toxic text. For a clearer comparison, we also categorize existing VLMs based on their safety levels and provide corresponding reinforcement recommendations. Our multimodal prompt dataset and red team code will be released after ethics committee approval. CONTENT WARNING: THIS PAPER CONTAINS HARMFUL MODEL RESPONSES., Comment: To be published in ACM MM 2024

Computer-aided design (CAD) tools are increasingly popular in modern dental practice, particularly for treatment planning or comprehensive prognosis evaluation. In particular, the 2D panoramic X-ray image efficiently detects invisible caries, impacted teeth and supernumerary teeth in children, while the 3D dental cone beam computed tomography (CBCT) is widely used in orthodontics and endodontics due to its low radiation dose. However, there is no open-access 2D public dataset for children's teeth and no open 3D dental CBCT dataset, which limits the development of automatic algorithms for segmenting teeth and analyzing diseases. The Semi-supervised Teeth Segmentation (STS) Challenge, a pioneering event in tooth segmentation, was held as a part of the MICCAI 2023 ToothFairy Workshop on the Alibaba Tianchi platform. This challenge aims to investigate effective semi-supervised tooth segmentation algorithms to advance the field of dentistry. In this challenge, we provide two modalities including the 2D panoramic X-ray images and the 3D CBCT tooth volumes. In Task 1, the goal was to segment tooth regions in panoramic X-ray images of both adult and pediatric teeth. Task 2 involved segmenting tooth sections using CBCT volumes. Limited labelled images with mostly unlabelled ones were provided in this challenge prompt using semi-supervised algorithms for training. In the preliminary round, the challenge received registration and result submission by 434 teams, with 64 advancing to the final round. This paper summarizes the diverse methods employed by the top-ranking teams in the STS MICCAI 2023 Challenge.

Security concerns for large language models (LLMs) have recently escalated, focusing on thwarting jailbreaking attempts in discrete prompts. However, the exploration of jailbreak vulnerabilities arising from continuous embeddings has been limited, as prior approaches primarily involved appending discrete or continuous suffixes to inputs. Our study presents a novel channel for conducting direct attacks on LLM inputs, eliminating the need for suffix addition or specific questions provided that the desired output is predefined. We additionally observe that extensive iterations often lead to overfitting, characterized by repetition in the output. To counteract this, we propose a simple yet effective strategy named CLIP. Our experiments show that for an input length of 40 at iteration 1000, applying CLIP improves the ASR from 62% to 83%

The proton drip line of (hyper)nuclei is examined within the framework of the deformed Skyrme-Hartree Fock approach by adjusting the nuclear force parameters to exactly reproduce the core binding energies. The impact of adding a {\Lambda} hyperon in a s or p state is studied, and it is found that in some cases the deformation effect facilitates the extension of the drip line by an added p-state hyperon. However, no extension of the drip line is found for s-state hypernuclei., Comment: 7 pages, 2 Figures

Large Language Models (LLMs) have showcased their remarkable capabilities in diverse domains, encompassing natural language understanding, translation, and even code generation. The potential for LLMs to generate harmful content is a significant concern. This risk necessitates rigorous testing and comprehensive evaluation of LLMs to ensure safe and responsible use. However, extensive testing of LLMs requires substantial computational resources, making it an expensive endeavor. Therefore, exploring cost-saving strategies during the testing phase is crucial to balance the need for thorough evaluation with the constraints of resource availability. To address this, our approach begins by transferring the moderation knowledge from an LLM to a small model. Subsequently, we deploy two distinct strategies for generating malicious queries: one based on a syntax tree approach, and the other leveraging an LLM-based method. Finally, our approach incorporates a sequential filter-test process designed to identify test cases that are prone to eliciting toxic responses. Our research evaluated the efficacy of DistillSeq across four LLMs: GPT-3.5, GPT-4.0, Vicuna-13B, and Llama-13B. In the absence of DistillSeq, the observed attack success rates on these LLMs stood at 31.5% for GPT-3.5, 21.4% for GPT-4.0, 28.3% for Vicuna-13B, and 30.9% for Llama-13B. However, upon the application of DistillSeq, these success rates notably increased to 58.5%, 50.7%, 52.5%, and 54.4%, respectively. This translated to an average escalation in attack success rate by a factor of 93.0% when compared to scenarios without the use of DistillSeq. Such findings highlight the significant enhancement DistillSeq offers in terms of reducing the time and resource investment required for effectively testing LLMs.

Predicting emotions elicited by news headlines can be challenging as the task is largely influenced by the varying nature of people's interpretations and backgrounds. Previous works have explored classifying discrete emotions directly from news headlines. We provide a different approach to tackling this problem by utilizing people's explanations of their emotion, written in free-text, on how they feel after reading a news headline. Using the dataset BU-NEmo+ (Gao et al., 2022), we found that for emotion classification, the free-text explanations have a strong correlation with the dominant emotion elicited by the headlines. The free-text explanations also contain more sentimental context than the news headlines alone and can serve as a better input to emotion classification models. Therefore, in this work we explored generating emotion explanations from headlines by training a sequence-to-sequence transformer model and by using pretrained large language model, ChatGPT (GPT-4). We then used the generated emotion explanations for emotion classification. In addition, we also experimented with training the pretrained T5 model for the intermediate task of explanation generation before fine-tuning it for emotion classification. Using McNemar's significance test, methods that incorporate GPT-generated free-text emotion explanations demonstrated significant improvement (P-value < 0.05) in emotion classification from headlines, compared to methods that only use headlines. This underscores the value of using intermediate free-text explanations for emotion prediction tasks with headlines., Comment: published at LREC-COLING 2024

We study CP violation (CPV) in the sneutrino sector within the B-L extension of the Minimal Supersymmetric Standard Model (BLSSM), wherein an inverse seesaw mechanism has been implemented. CPV arises from the new superpotential couplings in the (s)neutrino sector, which can be complex and the mixing of CP-eigenstates induced by those couplings. CPV leads to asymmetries in so called T-odd observables, but we argue that such asymmetries also lead to a wider distribution of those observables. We look at a final state where a sneutrino decays to a lepton, two jets and missing transverse momentum at the Future Circular Collider operating in hadron-hadron mode at $100$ TeV and with a luminosity of 3~ab$^{-1}$. In order to exclude the CP conserving scenario we need to improve traditional analysis by introducing boosted decision trees using both standard kinematic variables and T-odd observables and we need $Z^{\prime}$ boson not too much above current bounds as a portal to produce sneutrinos efficiently., Comment: 27 pages, 21 figures, 10 tables

Energy transfer between different mechanical degrees of freedom in atom-molecule collisions has been widely studied and largely understood. However, systems involving spins remain less explored, especially with a state-to-state precision. Here, we directly observed the energy transfer from atomic hyperfine to molecular rotation in the $^{87}$Rb ($|F_a,M_{F_a}\rangle = |2,2\rangle$) + $^{40}$K$^{87}$Rb (in the rovibronic ground state $N=0$) $\longrightarrow$ Rb ($ |1,1\rangle$) + KRb ($N=0,1,2$) exothermic collision. We probed the quantum states of the collision products using resonance-enhanced multi-photon ionization followed by time-of-flight mass spectrometry. We also carried out state-of-the-art quantum scattering calculations, which rigorously take into account the coupling between the spin and rotational degrees of freedom at short range, and assume that the KRb monomer can be treated as a rigid rotor moving on a single potential energy surface. The calculated product rotational state distribution deviates from the observations even after extensive tuning of the atom-molecule potential energy surface, suggesting that vibrational degrees of freedom and conical intersections play an important part in ultracold Rb + KRb collisions. Additionally, our ab initio calculations indicate that spin-rotation coupling is dramatically enhanced near a conical intersection, which is energetically accessible at short range. The observations confirm that spin is coupled to mechanical rotation at short range and establish a benchmark for future theoretical studies.

To support software developers in understanding and maintaining programs, various automatic (source) code summarization techniques have been proposed to generate a concise natural language summary (i.e., comment) for a given code snippet. Recently, the emergence of large language models (LLMs) has led to a great boost in the performance of code-related tasks. In this paper, we undertake a systematic and comprehensive study on code summarization in the era of LLMs, which covers multiple aspects involved in the workflow of LLM-based code summarization. Specifically, we begin by examining prevalent automated evaluation methods for assessing the quality of summaries generated by LLMs and find that the results of the GPT-4 evaluation method are most closely aligned with human evaluation. Then, we explore the effectiveness of five prompting techniques (zero-shot, few-shot, chain-of-thought, critique, and expert) in adapting LLMs to code summarization tasks. Contrary to expectations, advanced prompting techniques may not outperform simple zero-shot prompting. Next, we investigate the impact of LLMs' model settings (including top\_p and temperature parameters) on the quality of generated summaries. We find the impact of the two parameters on summary quality varies by the base LLM and programming language, but their impacts are similar. Moreover, we canvass LLMs' abilities to summarize code snippets in distinct types of programming languages. The results reveal that LLMs perform suboptimally when summarizing code written in logic programming languages compared to other language types. Finally, we unexpectedly find that CodeLlama-Instruct with 7B parameters can outperform advanced GPT-4 in generating summaries describing code implementation details and asserting code properties. We hope that our findings can provide a comprehensive understanding of code summarization in the era of LLMs., Comment: Just accepted to the 47th International Conference on Software Engineering (ICSE 2025)

We revisit the question of what randomization ratio (RR) maximizes power of the logrank test in event-driven survival trials under proportional hazards (PH). By comparing three approximations of the logrank test (Schoenfeld, Freedman, Rubinstein) to empirical simulations, we find that the RR that maximizes power is the RR that balances number of events across treatment arms at the end of the trial. This contradicts the common misconception implied by Schoenfeld's approximation that 1:1 randomization maximizes power. Besides power, we consider other factors that might influence the choice of RR (accrual, trial duration, sample size, etc.). We perform simulations to better understand how unequal randomization might impact these factors in practice. Altogether, we derive 6 insights to guide statisticians in the design of survival trials considering unequal randomization., Comment: 17 pages, 3 figures, 2 tables

In this paper, we analyze the spectral energy distributions (SEDs) of 17 powerful (with a spin-down luminosity greater than $10^{35}$ erg s$^{-1}$) young (with an age less than 15000 yrs) pulsar wind nebulae (PWNe) using a simple time-independent one-zone emission model. Our aim is to investigate correlations between model parameters and the ages of the corresponding PWNe, thereby revealing the evolution of high-energy electron distributions within PWNe. Our findings are as follows: (1) The electron distributions in PWNe can be characterized by a double power-law with a superexponential cutoff; (2) As PWNe evolve, the high-energy end of the electron distribution spectrum becomes harder with the index decreasing from approximately 3.5 to 2.5, while the low-energy end spectrum index remains constant near 1.5; (3) There is no apparent correlation between the break energy or cutoff energy and the age of PWNe. (4) The average magnetic field within PWNe decreases with age, leading to a positive correlation between the energy loss timescale of electrons at the break energy or the high-energy cutoff, and the age of the PWN. (5) The total electron energy within PWNe remains constant near $2 \times 10^{48}$ erg, while the total magnetic energy decreases with age., Comment: 19 pages, 20 figures, 1 table accepted for publication in RAA

Large Language Models (LLMs) have shown superior performance in various applications and fields. To achieve better performance on specialized domains such as law and advertisement, LLMs are often continue pre-trained on in-domain data. However, existing approaches suffer from two major issues. First, in-domain data are scarce compared with general domain-agnostic data. Second, data used for continual pre-training are not task-aware, such that they may not be helpful to downstream applications. We propose TRAIT, a task-oriented in-domain data augmentation framework. Our framework is divided into two parts: in-domain data selection and task-oriented synthetic passage generation. The data selection strategy identifies and selects a large amount of in-domain data from general corpora, and thus significantly enriches domain knowledge in the continual pre-training data. The synthetic passages contain guidance on how to use domain knowledge to answer questions about downstream tasks. By training on such passages, the model aligns with the need of downstream applications. We adapt LLMs to two domains: advertisement and math. On average, TRAIT improves LLM performance by 8% in the advertisement domain and 7.5% in the math domain.

Text-to-Image Diffusion Models (T2I DMs) have garnered significant attention for their ability to generate high-quality images from textual descriptions. However, these models often produce images that do not fully align with the input prompts, resulting in semantic inconsistencies. The most prominent issue among these semantic inconsistencies is catastrophic-neglect, where the images generated by T2I DMs miss key objects mentioned in the prompt. We first conduct an empirical study on this issue, exploring the prevalence of catastrophic-neglect, potential mitigation strategies with feature enhancement, and the insights gained. Guided by the empirical findings, we propose an automated repair approach named Patcher to address catastrophic-neglect in T2I DMs. Specifically, Patcher first determines whether there are any neglected objects in the prompt, and then applies attention-guided feature enhancement to these neglected objects, resulting in a repaired prompt. Experimental results on three versions of Stable Diffusion demonstrate that Patcher effectively repairs the issue of catastrophic-neglect, achieving 10.1%-16.3% higher Correct Rate in image generation compared to baselines., Comment: 11 pages, 3 figures

Federated Learning (FL) is susceptible to poisoning attacks, wherein compromised clients manipulate the global model by modifying local datasets or sending manipulated model updates. Experienced defenders can readily detect and mitigate the poisoning effects of malicious behaviors using Byzantine-robust aggregation rules. However, the exploration of poisoning attacks in scenarios where such behaviors are absent remains largely unexplored for Byzantine-robust FL. This paper addresses the challenging problem of poisoning Byzantine-robust FL by introducing catastrophic forgetting. To fill this gap, we first formally define generalization error and establish its connection to catastrophic forgetting, paving the way for the development of a clean-label data poisoning attack named BadSampler. This attack leverages only clean-label data (i.e., without poisoned data) to poison Byzantine-robust FL and requires the adversary to selectively sample training data with high loss to feed model training and maximize the model's generalization error. We formulate the attack as an optimization problem and present two elegant adversarial sampling strategies, Top-$\kappa$ sampling, and meta-sampling, to approximately solve it. Additionally, our formal error upper bound and time complexity analysis demonstrate that our design can preserve attack utility with high efficiency. Extensive evaluations on two real-world datasets illustrate the effectiveness and performance of our proposed attacks., Comment: In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD' 24), August 25-29, 2024, Barcelona, Spain

Large Language Models (LLMs) like OpenAI's GPT series, Anthropic's Claude, and Meta's LLaMa have shown remarkable capabilities in text generation. However, their susceptibility to toxic prompts presents significant security challenges. This paper investigates alignment techniques, including Supervised Fine-Tuning (SFT) and Reinforcement Learning from Human Feedback (RLHF), to mitigate these risks. We conduct an empirical study on refusal patterns across nine LLMs, revealing that models with uniform refusal patterns, such as Claude3, exhibit higher security. Based on these findings, we propose self-distilling and cross-model distilling methods to enhance LLM security. Our results show that these methods significantly improve refusal rates and reduce unsafe content, with cross-model distilling achieving refusal rates close to Claude3's 94.51%. These findings underscore the potential of distillation-based alignment in securing LLMs against toxic prompts.

We present a novel benchmark dataset and prediction tasks for investigating approaches to assess cognitive function through analysis of connected speech. The dataset consists of speech samples and clinical information for speakers of Mandarin Chinese and English with different levels of cognitive impairment as well as individuals with normal cognition. These data have been carefully matched by age and sex by propensity score analysis to ensure balance and representativity in model training. The prediction tasks encompass mild cognitive impairment diagnosis and cognitive test score prediction. This framework was designed to encourage the development of approaches to speech-based cognitive assessment which generalise across languages. We illustrate it by presenting baseline prediction models that employ language-agnostic and comparable features for diagnosis and cognitive test score prediction. The models achieved unweighted average recall was 59.2% in diagnosis, and root mean squared error of 2.89 in score prediction., Comment: To appear in Proceedings of Interspeech 2024

Simulations have played a critical role in the advancement of our knowledge of magnetic reconnection. However, due to the inherently multiscale nature of reconnection, it is impossible to simulate all physics at all scales. For this reason, a wide range of simulation methods have been crafted to study particular aspects and consequences of magnetic reconnection. This chapter reviews many of these methods, laying out critical assumptions, numerical techniques, and giving examples of scientific results. Plasma models described include magnetohydrodynamics (MHD), Hall MHD, Hybrid, kinetic particle-in-cell (PIC), kinetic Vlasov, Fluid models with embedded PIC, Fluid models with direct feedback from energetic populations, and the Rice Convection Model (RCM)., Comment: Chapter 5.2 of ISSI Book on Magnetic Reconnection, submitted to Space Science Reviews

Materials that exhibit strongly coupled magnetic order and electronic properties are crucial for both fundamental research and technological applications. However, finding a material that not only shows remarkable magnetoresistive responses but also has an easily tunable ground state remains a challenge. Here, we report successful manipulation of the magnetic and transport properties of EuCd$_2$P$_2$, which is transformed from an A-type antiferromagnet ($T_\mathrm{N}$ = 11 K) exhibiting colossal magnetoresistance into a ferromagnet ($T_\mathrm{C}$ = 47 K) with metallic behavior. The dramatic alteration results from a low hole concentration of $10^{19}$ cm$^{-3}$ induced by changing the growth conditions. Electronic structure and total energy calculations confirm the tunability of magnetism with a small carrier concentration for EuCd$_2$P$_2$. It is feasible to switch between the magnetic states by using field-effect to control the carrier density, thereby changing the magneto-electronic response. The controllable magnetism and electrical transport of EuCd$_2$P$_2$ make it a potential candidate for spintronics.

The pyroresistive response of conductive polymer composites (CPCs) has attracted much interest because of its potential applications in many electronic devices requiring a significant responsiveness to changes in external physical parameters such as temperature or electric fields. Although extensive research has been conducted to study how the properties of the polymeric matrix and conductive fillers affect the positive temperature coefficient pyroresistive effect, the understanding of the microscopic mechanism governing such a phenomenon is still incomplete. In particular, to date, there is little body of theoretical research devoted to investigating the effect of the polymer thermal expansion on the electrical connectivity of the conductive phase. Here, we present the results of simulations of model CPCs in which rigid conductive fillers are dispersed in an insulating amorphous matrix. By employing a meshless algorithm to analyze the thermoelastic response of the system, we couple the computed strain field to the electrical connectedness of the percolating conductive particles. We show that the electrical conductivity responds to the local strains that are generated by the mismatch between the thermal expansion of the polymeric and conductive phases and that the conductor-insulator transition is caused by a sudden and global disconnection of the electrical contacts forming the percolating network., Comment: 9 pages with 7 figures: supplemental material: 7 pages with 6 figures

Obstructed atomic insulator is recently proposed as an unconventional material, in which electric charge centers localized at sites away from the atoms. A half-filling surface state would emerge at specific interfaces cutting through these charge centers and avoid intersecting any atoms. In this article, we utilized angle-resolved photoemission spectroscopy and density functional theory calculations to study one of the obstructed atomic insulator candidates, NiP$_2$. A floating surface state with large effective mass that is isolated from all bulk states is resolved on the (100) cleavage plane, distinct from previously reported surface states in obstructed atomic insulators that are merged into bulk bands. Density functional theory calculation results elucidate that this floating surface state is originated from the obstructed Wannier charge centers, albeit underwent surface reconstruction that splits the half-filled obstructed surface state. Our findings not only shed lights on the spectroscopy study of obstructed atomic insulators and obstructed surface states, but also provide possible route for development of new catalysts., Comment: 21 pages, 5 figures

Kagome materials exhibit many novel phenomena emerging from the interplay between lattice geometry, electronic structure, and topology. A prime example is the vanadium-based kagome materials AV3Sb5 (A = K, Rb, and Cs) with superconductivity and unconventional charge-density wave (CDW). More interestingly, the substitution of vanadium by chromium further introduces magnetism and enhances the correlation effect in CsCr3Sb5 which likewise exhibits superconductivity under pressure and competing density-wave state. Here we systematically investigate the electronic structure of CsCr3Sb5 using high-resolution angle-resolved photoemission spectroscopy (APRES) and ab-initio calculations. Overall, the measured electronic structure agrees with the theoretical calculation. Remarkably, Cr 3d orbitals exhibit incoherent electronic states and contribute to incipient flat bands close to the Fermi level. The electronic structure shows a minor change across the magnetic transition at 55 K, suggesting a weak interplay between the local magnetic moment and itinerant electrons. Furthermore, we reveal a drastic enhancement of the electron scattering rate across the magnetic transition, which is relevant to the semiconducting-like transport property of the system at high temperatures. Our results suggest that CsCr3Sb5 is a strongly correlated Hund's metal with incipient flat bands near the Fermi level, which provides an electronic basis for understanding its novel properties in comparison to the non-magnetic and weakly correlated AV3Sb5.

Magnetic reconnection is a ubiquitous plasma process that transforms magnetic energy into particle energy during eruptive events throughout the universe. Reconnection not only converts energy during solar flares and geomagnetic substorms that drive space weather near Earth, but it may also play critical roles in the high energy emissions from the magnetospheres of neutron stars and black holes. In this review article, we focus on collisionless plasmas that are most relevant to reconnection in many space and astrophysical plasmas. Guided by first-principles kinetic simulations and spaceborne in-situ observations, we highlight the most recent progress in understanding this fundamental plasma process. We start by discussing the non-ideal electric field in the generalized Ohm's law that breaks the frozen-in flux condition in ideal magnetohydrodynamics and allows magnetic reconnection to occur. We point out that this same reconnection electric field also plays an important role in sustaining the current and pressure in the current sheet and then discuss the determination of its magnitude (i.e., the reconnection rate), based on force balance and energy conservation. This approach to determining the reconnection rate is applied to kinetic current sheets of a wide variety of magnetic geometries, parameters, and background conditions. We also briefly review the key diagnostics and modeling of energy conversion around the reconnection diffusion region, seeking insights from recently developed theories. Finally, future prospects and open questions are discussed., Comment: Submitted to Space Science Reviews. This is a review paper as an outcome of the 2022 Magnetic Reconnection Workshop in the International Space Science Institute

Multi-aspect controllable text generation aims to control the generated texts in attributes from multiple aspects (e.g., "positive" from sentiment and "sport" from topic). For ease of obtaining training samples, existing works neglect attribute correlations formed by the intertwining of different attributes. Particularly, the stereotype formed by imbalanced attribute correlations significantly affects multi-aspect control. In this paper, we propose MAGIC, a new multi-aspect controllable text generation method with disentangled counterfactual augmentation. We alleviate the issue of imbalanced attribute correlations during training using counterfactual feature vectors in the attribute latent space by disentanglement. During inference, we enhance attribute correlations by target-guided counterfactual augmentation to further improve multi-aspect control. Experiments show that MAGIC outperforms state-of-the-art baselines in both imbalanced and balanced attribute correlation scenarios. Our source code and data are available at https://github.com/nju-websoft/MAGIC., Comment: Accepted in the 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024)

Video text spotting (VTS) aims to simultaneously localize, recognize and track text instances in videos. To address the limited recognition capability of end-to-end methods, recent methods track the zero-shot results of state-of-the-art image text spotters directly, and achieve impressive performance. However, owing to the domain gap between different datasets, these methods usually obtain limited tracking trajectories on extreme dataset. Fine-tuning transformer-based text spotters on specific datasets could yield performance enhancements, albeit at the expense of considerable training resources. In this paper, we propose a Language Collaboration and Glyph Perception Model, termed LOGO, an innovative framework designed to enhance the performance of conventional text spotters. To achieve this goal, we design a language synergy classifier (LSC) to explicitly discern text instances from background noise in the recognition stage. Specially, the language synergy classifier can output text content or background code based on the legibility of text regions, thus computing language scores. Subsequently, fusion scores are computed by taking the average of detection scores and language scores, and are utilized to re-score the detection results before tracking. By the re-scoring mechanism, the proposed LSC facilitates the detection of low-resolution text instances while filtering out text-like regions. Moreover, the glyph supervision is introduced to enhance the recognition accuracy of noisy text regions. In addition, we propose the visual position mixture module, which can merge the position information and visual features efficiently, and acquire more discriminative tracking features. Extensive experiments on public benchmarks validate the effectiveness of the proposed method.

Recently, tremendous strides have been made to align the generation of Large Language Models (LLMs) with human values to mitigate toxic or unhelpful content. Leveraging Reinforcement Learning from Human Feedback (RLHF) proves effective and is widely adopted by researchers. However, implementing RLHF is complex, and its sensitivity to hyperparameters renders achieving stable performance and scalability challenging. Furthermore, prevailing approaches to preference alignment primarily concentrate on pairwise comparisons, with limited exploration into multi-response scenarios, thereby overlooking the potential richness within the candidate pool. For the above reasons, we propose a new approach: Listwise Reward Enhancement for Preference Alignment (LIRE), a gradient-based reward optimization approach that incorporates the offline rewards of multiple responses into a streamlined listwise framework, thus eliminating the need for online sampling during training. LIRE is straightforward to implement, requiring minimal parameter tuning, and seamlessly aligns with the pairwise paradigm while naturally extending to multi-response scenarios. Moreover, we introduce a self-enhancement algorithm aimed at iteratively refining the reward during training. Our experiments demonstrate that LIRE consistently outperforms existing methods across several benchmarks on dialogue and summarization tasks, with good transferability to out-of-distribution data, assessed using proxy reward models and human annotators., Comment: Accepted by ACL 2024 Findings

Large language models (LLMs) have transformed the field of natural language processing, but they remain susceptible to jailbreaking attacks that exploit their capabilities to generate unintended and potentially harmful content. Existing token-level jailbreaking techniques, while effective, face scalability and efficiency challenges, especially as models undergo frequent updates and incorporate advanced defensive measures. In this paper, we introduce JailMine, an innovative token-level manipulation approach that addresses these limitations effectively. JailMine employs an automated "mining" process to elicit malicious responses from LLMs by strategically selecting affirmative outputs and iteratively reducing the likelihood of rejection. Through rigorous testing across multiple well-known LLMs and datasets, we demonstrate JailMine's effectiveness and efficiency, achieving a significant average reduction of 86% in time consumed while maintaining high success rates averaging 95%, even in the face of evolving defensive strategies. Our work contributes to the ongoing effort to assess and mitigate the vulnerability of LLMs to jailbreaking attacks, underscoring the importance of continued vigilance and proactive measures to enhance the security and reliability of these powerful language models.

Recent years have experienced increasing utilization of complex machine learning models across multiple sources of data to inform more generalizable decision-making. However, distribution shifts across data sources and privacy concerns related to sharing individual-level data, coupled with a lack of uncertainty quantification from machine learning predictions, make it challenging to achieve valid inferences in multi-source environments. In this paper, we consider the problem of obtaining distribution-free prediction intervals for a target population, leveraging multiple potentially biased data sources. We derive the efficient influence functions for the quantiles of unobserved outcomes in the target and source populations, and show that one can incorporate machine learning prediction algorithms in the estimation of nuisance functions while still achieving parametric rates of convergence to nominal coverage probabilities. Moreover, when conditional outcome invariance is violated, we propose a data-adaptive strategy to upweight informative data sources for efficiency gain and downweight non-informative data sources for bias reduction. We highlight the robustness and efficiency of our proposals for a variety of conformal scores and data-generating mechanisms via extensive synthetic experiments. Hospital length of stay prediction intervals for pediatric patients undergoing a high-risk cardiac surgical procedure between 2016-2022 in the U.S. illustrate the utility of our methodology., Comment: Accepted to ICML 2024, 39 pages, 13 figures

Pathology is the study of microscopic inspection of tissue, and a pathology diagnosis is often the medical gold standard to diagnose disease. Pathology images provide a unique challenge for computer-vision-based analysis: a single pathology Whole Slide Image (WSI) is gigapixel-sized and often contains hundreds of thousands to millions of objects of interest across multiple resolutions. In this work, we propose PathoLogy Universal TransfOrmer (PLUTO): a light-weight pathology FM that is pre-trained on a diverse dataset of 195 million image tiles collected from multiple sites and extracts meaningful representations across multiple WSI scales that enable a large variety of downstream pathology tasks. In particular, we design task-specific adaptation heads that utilize PLUTO's output embeddings for tasks which span pathology scales ranging from subcellular to slide-scale, including instance segmentation, tile classification, and slide-level prediction. We compare PLUTO's performance to other state-of-the-art methods on a diverse set of external and internal benchmarks covering multiple biologically relevant tasks, tissue types, resolutions, stains, and scanners. We find that PLUTO matches or outperforms existing task-specific baselines and pathology-specific foundation models, some of which use orders-of-magnitude larger datasets and model sizes when compared to PLUTO. Our findings present a path towards a universal embedding to power pathology image analysis, and motivate further exploration around pathology foundation models in terms of data diversity, architectural improvements, sample efficiency, and practical deployability in real-world applications.

CAPTCHAs have become a ubiquitous tool in safeguarding applications from automated bots. Over time, the arms race between CAPTCHA development and evasion techniques has led to increasingly sophisticated and diverse designs. The latest iteration, reasoning CAPTCHAs, exploits tasks that are intuitively simple for humans but challenging for conventional AI technologies, thereby enhancing security measures. Driven by the evolving AI capabilities, particularly the advancements in Large Language Models (LLMs), we investigate the potential of multimodal LLMs to solve modern reasoning CAPTCHAs. Our empirical analysis reveals that, despite their advanced reasoning capabilities, LLMs struggle to solve these CAPTCHAs effectively. In response, we introduce Oedipus, an innovative end-to-end framework for automated reasoning CAPTCHA solving. Central to this framework is a novel strategy that dissects the complex and human-easy-AI-hard tasks into a sequence of simpler and AI-easy steps. This is achieved through the development of a Domain Specific Language (DSL) for CAPTCHAs that guides LLMs in generating actionable sub-steps for each CAPTCHA challenge. The DSL is customized to ensure that each unit operation is a highly solvable subtask revealed in our previous empirical study. These sub-steps are then tackled sequentially using the Chain-of-Thought (CoT) methodology. Our evaluation shows that Oedipus effectively resolves the studied CAPTCHAs, achieving an average success rate of 63.5\%. Remarkably, it also shows adaptability to the most recent CAPTCHA designs introduced in late 2023, which are not included in our initial study. This prompts a discussion on future strategies for designing reasoning CAPTCHAs that can effectively counter advanced AI solutions.

We present a scene representation, which we call a tactile-augmented radiance field (TaRF), that brings vision and touch into a shared 3D space. This representation can be used to estimate the visual and tactile signals for a given 3D position within a scene. We capture a scene's TaRF from a collection of photos and sparsely sampled touch probes. Our approach makes use of two insights: (i) common vision-based touch sensors are built on ordinary cameras and thus can be registered to images using methods from multi-view geometry, and (ii) visually and structurally similar regions of a scene share the same tactile features. We use these insights to register touch signals to a captured visual scene, and to train a conditional diffusion model that, provided with an RGB-D image rendered from a neural radiance field, generates its corresponding tactile signal. To evaluate our approach, we collect a dataset of TaRFs. This dataset contains more touch samples than previous real-world datasets, and it provides spatially aligned visual signals for each captured touch signal. We demonstrate the accuracy of our cross-modal generative model and the utility of the captured visual-tactile data on several downstream tasks. Project page: https://dou-yiming.github.io/TaRF, Comment: CVPR 2024, Project page: https://dou-yiming.github.io/TaRF, Code: https://github.com/Dou-Yiming/TaRF/

Transition metal compounds with kagome structure have been found to exhibit a variety of exotic structural, electronic, and magnetic orders. These orders are competing with energies very close to each other, resulting in complex phase transitions. Some of the phases are easily observable, such as the charge density wave (CDW) and the superconducting phase, while others are more challenging to identify and characterize. Here we present magneto-transport evidence of a new phase below ~35 K in the kagome topological metal CsV$_3$Sb$_5$ (CVS) thin flakes between the CDW and the superconducting transition temperatures. This phase is characterized by six-fold rotational symmetry in the in-plane magnetoresistance (MR) and is connected to the orbital current order in CVS. Furthermore, the phase is characterized by a large in-plane negative magnetoresistance, which suggests the existence of a three-dimensional, magnetic field-tunable orbital current ordered phase. Our results highlight the potential of magneto-transport to reveal the interactions between exotic quantum states of matter and to uncover the symmetry of such hidden phases.

Federated learning (FL) has attracted widespread attention because it supports the joint training of models by multiple participants without moving private dataset. However, there are still many security issues in FL that deserve discussion. In this paper, we consider three major issues: 1) how to ensure that the training process can be publicly audited by any third party; 2) how to avoid the influence of malicious participants on training; 3) how to ensure that private gradients and models are not leaked to third parties. Many solutions have been proposed to address these issues, while solving the above three problems simultaneously is seldom considered. In this paper, we propose a publicly auditable and privacy-preserving federated learning scheme that is resistant to malicious participants uploading gradients with wrong directions and enables anyone to audit and verify the correctness of the training process. In particular, we design a robust aggregation algorithm capable of detecting gradients with wrong directions from malicious participants. Then, we design a random vector generation algorithm and combine it with zero sharing and blockchain technologies to make the joint training process publicly auditable, meaning anyone can verify the correctness of the training. Finally, we conduct a series of experiments, and the experimental results show that the model generated by the protocol is comparable in accuracy to the original FL approach while keeping security advantages., Comment: ICC 2024 - 2024 IEEE International Conference on Communications Conference Program

We propose a theoretical scheme for a non-Hermitian atomic grating within an ultra-cold rubidium-87 ($^{87}Rb$) atomic ensemble. The grating's diffraction properties depend on the polarization states of incident photons and are controlled non-locally through Rydberg interactions. Multiple types of polarization-dependent diffraction modes are generated, benefiting from no crosstalk atomic transition channels based on transition selection rules. Those polarization-dependent diffraction modes can be switched using dynamic optical pulse trains, exploiting the Rydberg blockade effect, and are tunable by non-Hermitian optical modulation. Our work will advance the application of asymmetric optical scattering by utilizing the polarization degree of freedom within continuous media and benefit the application of versatile non-Hermitian/asymmetric optical devices.

Large language models (LLMs) have transformed the landscape of language processing, yet struggle with significant challenges in terms of security, privacy, and the generation of seemingly coherent but factually inaccurate outputs, commonly referred to as hallucinations. Among these challenges, one particularly pressing issue is Fact-Conflicting Hallucination (FCH), where LLMs generate content that directly contradicts established facts. Tackling FCH poses a formidable task due to two primary obstacles: Firstly, automating the construction and updating of benchmark datasets is challenging, as current methods rely on static benchmarks that don't cover the diverse range of FCH scenarios. Secondly, validating LLM outputs' reasoning process is inherently complex, especially with intricate logical relations involved. In addressing these obstacles, we propose an innovative approach leveraging logic programming to enhance metamorphic testing for detecting Fact-Conflicting Hallucinations (FCH). Our method gathers data from sources like Wikipedia, expands it with logical reasoning to create diverse test cases, assesses LLMs through structured prompts, and validates their coherence using semantic-aware assessment mechanisms. Our method generates test cases and detects hallucinations across six different LLMs spanning nine domains, revealing hallucination rates ranging from 24.7% to 59.8%. Key observations indicate that LLMs encounter challenges, particularly with temporal concepts, handling out-of-distribution knowledge, and exhibiting deficiencies in logical reasoning capabilities. The outcomes underscore the efficacy of logic-based test cases generated by our tool in both triggering and identifying hallucinations. These findings underscore the imperative for ongoing collaborative endeavors within the community to detect and address LLM hallucinations.

Adverse reactions caused by drug combinations are an increasingly common phenomenon, making their accurate prediction an important challenge in modern medicine. However, the polynomial nature of this problem renders lab-based identification of adverse reactions insufficient. Dozens of computational approaches have therefore been proposed for the task in recent years, with varying degrees of success. One group of methods that has seemingly been under-utilised in this area is tensor factorisation, despite their clear applicability to this type of data. In this work, we apply three such models to a benchmark dataset in order to compare them against established techniques. We find, in contrast to previous reports, that for this task tensor factorisation models are competitive with state-of-the-art graph neural network models and we recommend that future work in this field considers cheaper methods with linear complexity before running costly deep learning processes.

This paper provides a comprehensive review of the NTIRE 2024 challenge, focusing on efficient single-image super-resolution (ESR) solutions and their outcomes. The task of this challenge is to super-resolve an input image with a magnification factor of x4 based on pairs of low and corresponding high-resolution images. The primary objective is to develop networks that optimize various aspects such as runtime, parameters, and FLOPs, while still maintaining a peak signal-to-noise ratio (PSNR) of approximately 26.90 dB on the DIV2K_LSDIR_valid dataset and 26.99 dB on the DIV2K_LSDIR_test dataset. In addition, this challenge has 4 tracks including the main track (overall performance), sub-track 1 (runtime), sub-track 2 (FLOPs), and sub-track 3 (parameters). In the main track, all three metrics (ie runtime, FLOPs, and parameter count) were considered. The ranking of the main track is calculated based on a weighted sum-up of the scores of all other sub-tracks. In sub-track 1, the practical runtime performance of the submissions was evaluated, and the corresponding score was used to determine the ranking. In sub-track 2, the number of FLOPs was considered. The score calculated based on the corresponding FLOPs was used to determine the ranking. In sub-track 3, the number of parameters was considered. The score calculated based on the corresponding parameters was used to determine the ranking. RLFN is set as the baseline for efficiency measurement. The challenge had 262 registered participants, and 34 teams made valid submissions. They gauge the state-of-the-art in efficient single-image super-resolution. To facilitate the reproducibility of the challenge and enable other researchers to build upon these findings, the code and the pre-trained model of validated solutions are made publicly available at https://github.com/Amazingren/NTIRE2024_ESR/., Comment: The report paper of NTIRE2024 Efficient Super-resolution, accepted by CVPRW2024

Understanding how the structure of materials affects their properties is a cornerstone of materials science and engineering. However, traditional methods have struggled to accurately describe the quantitative structure-property relationships for complex structures. In our study, we bridge this gap by leveraging machine learning to analyze images of materials' microstructures, thus offering a novel way to understand and predict the properties of materials based on their microstructures. We introduce a method known as FAGC (Feature Augmentation on Geodesic Curves), specifically demonstrated for Cu-Cr-Zr alloys. This approach utilizes machine learning to examine the shapes within images of the alloys' microstructures and predict their mechanical and electronic properties. This generative FAGC approach can effectively expand the relatively small training datasets due to the limited availability of materials images labeled with quantitative properties. The process begins with extracting features from the images using neural networks. These features are then mapped onto the Pre-shape space to construct the Geodesic curves. Along these curves, new features are generated, effectively increasing the dataset. Moreover, we design a pseudo-labeling mechanism for these newly generated features to further enhance the training dataset. Our FAGC method has shown remarkable results, significantly improving the accuracy of predicting the electronic conductivity and hardness of Cu-Cr-Zr alloys, with R-squared values of 0.978 and 0.998, respectively. These outcomes underscore the potential of FAGC to address the challenge of limited image data in materials science, providing a powerful tool for establishing detailed and quantitative relationships between complex microstructures and material properties.

With the expanding application of Large Language Models (LLMs) in various domains, it becomes imperative to comprehensively investigate their unforeseen behaviors and consequent outcomes. In this study, we introduce and systematically explore the phenomenon of "glitch tokens", which are anomalous tokens produced by established tokenizers and could potentially compromise the models' quality of response. Specifically, we experiment on seven top popular LLMs utilizing three distinct tokenizers and involving a totally of 182,517 tokens. We present categorizations of the identified glitch tokens and symptoms exhibited by LLMs when interacting with glitch tokens. Based on our observation that glitch tokens tend to cluster in the embedding space, we propose GlitchHunter, a novel iterative clustering-based technique, for efficient glitch token detection. The evaluation shows that our approach notably outperforms three baseline methods on eight open-source LLMs. To the best of our knowledge, we present the first comprehensive study on glitch tokens. Our new detection further provides valuable insights into mitigating tokenization-related errors in LLMs.