llm - 2025_03
Navigation
Papers
The rapid development of large language models (LLMs) gives rise to ethical concerns about their performance, while opening new avenues for developing toxic language detection techniques. However, LLMs' unethical output and their capability of detecting toxicity have primarily been tested on language data that do not demand complex meaning inference, such as the biased associations of 'he' with programmer and 'she' with household. Nowadays toxic language adopts a much more creative range of implicit forms, thanks to advanced censorship. In this study, we collect authentic toxic interactions that evade online censorship and that are verified by human annotators as inference intensive. To evaluate and improve LLMs' reasoning of the authentic implicit toxic language, we propose a new prompting method, Pragmatic Inference Chain (PIC), drawn on interdisciplinary findings from cognitive science and linguistics. The PIC prompting significantly improves the success rate of GPT-4o, Llama-3.1-70B-Instruct, and DeepSeek-v2.5 in identifying implicit toxic language, compared to both direct prompting and Chain-of-Thought. In addition, it also facilitates the models to produce more explicit and coherent reasoning processes, hence can potentially be generalized to other inference-intensive tasks, e.g., understanding humour and metaphors.
Large language models (LLMs) have demonstrated remarkable capabilities in simulating human behaviour and social intelligence. However, they risk perpetuating societal biases, especially when demographic information is involved. We introduce a novel framework using cosine distance to measure semantic shifts in responses and an LLM-judged Preference Win Rate (WR) to assess how demographic prompts affect response quality across power-disparate social scenarios. Evaluating five LLMs over 100 diverse social scenarios and nine demographic axes, our findings suggest a "default persona" bias toward middle-aged, able-bodied, native-born, Caucasian, atheistic males with centrist views. Moreover, interactions involving specific demographics are associated with lower-quality responses. Lastly, the presence of power disparities increases variability in response semantics and quality across demographic groups, suggesting that implicit biases may be heightened under power-imbalanced conditions. These insights expose the demographic biases inherent in LLMs and offer potential paths toward future bias mitigation efforts in LLMs.
We present a survey of methods for assessing and enhancing the quality of online discussions, focusing on the potential of Large Language Models (LLMs). While online discourses aim, at least in theory, to foster mutual understanding, they often devolve into harmful exchanges, such as hate speech, threatening social cohesion and democratic values. Recent advancements in LLMs enable facilitation agents that not only moderate content, but also actively improve the quality of interactions. Our survey synthesizes ideas from Natural Language Processing (NLP) and Social Sciences to provide (a) a new taxonomy on discussion quality evaluation, (b) an overview of intervention and facilitation strategies, along with a new taxonomy on conversation facilitation datasets, (c) an LLM-oriented roadmap of good practices and future research directions, from technological and societal perspectives.
Large language models (LLMs) have demonstrated remarkable capabilities in tool learning. In real-world scenarios, user queries are often ambiguous and incomplete, requiring effective clarification. However, existing interactive clarification approaches face two critical limitations: reliance on manually constructed datasets and lack of error correction mechanisms during multi-turn clarification. We present AskToAct, which addresses these challenges by exploiting the structural mapping between queries and their tool invocation solutions. Our key insight is that tool parameters naturally represent explicit user intents. By systematically removing key parameters from queries while retaining them as ground truth, we enable automated construction of high-quality training data. We further enhance model robustness by fine-tuning on error-correction augmented data using selective masking mechanism, enabling dynamic error detection during clarification interactions. Comprehensive experiments demonstrate that AskToAct significantly outperforms existing approaches, achieving above 79% accuracy in recovering critical unspecified intents and enhancing clarification efficiency by an average of 48.34% while maintaining high accuracy in tool invocation. Our framework exhibits robust performance across varying complexity levels and successfully generalizes to entirely unseen APIs without additional training, achieving performance comparable to GPT-4 with substantially fewer computational resources.
One of the challenges of quantizing a large language model (LLM) is the presence of outliers. Outliers often make uniform quantization schemes less effective, particularly in extreme cases such as 4-bit quantization. We introduce KurTail, a new post-training quantization (PTQ) scheme that leverages Kurtosis-based rotation to mitigate outliers in the activations of LLMs. Our method optimizes Kurtosis as a measure of tailedness. This approach enables the quantization of weights, activations, and the KV cache in 4 bits. We utilize layer-wise optimization, ensuring memory efficiency. KurTail outperforms existing quantization methods, offering a 13.3\% boost in MMLU accuracy and a 15.5\% drop in Wiki perplexity compared to QuaRot. It also outperforms SpinQuant with a 2.6\% MMLU gain and reduces perplexity by 2.9\%, all while reducing the training cost. For comparison, learning the rotation using SpinQuant for Llama3-70B requires at least four NVIDIA H100 80GB GPUs, whereas our method requires only a single GPU, making it a more accessible solution for consumer GPU.
Textual data from social platforms captures various aspects of mental health through discussions around and across issues, while users reach out for help and others sympathize and offer support. We propose a comprehensive framework that leverages Natural Language Processing (NLP) and Generative AI techniques to identify and assess mental health disorders, detect their severity, and create recommendations for behavior change and therapeutic interventions based on users' posts on Reddit. To classify the disorders, we use rule-based labeling methods as well as advanced pre-trained NLP models to extract nuanced semantic features from the data. We fine-tune domain-adapted and generic pre-trained NLP models based on predictions from specialized Large Language Models (LLMs) to improve classification accuracy. Our hybrid approach combines the generalization capabilities of pre-trained models with the domain-specific insights captured by LLMs, providing an improved understanding of mental health discourse. Our findings highlight the strengths and limitations of each model, offering valuable insights into their practical applicability. This research potentially facilitates early detection and personalized care to aid practitioners and aims to facilitate timely interventions and improve overall well-being, thereby contributing to the broader field of mental health surveillance and digital health analytics.
Owing to the exceptional performance of Large Language Models (LLMs) in Natural Language Processing (NLP) tasks, LLM-based NLP software has rapidly gained traction across various domains, such as financial analysis and content moderation. However, these applications frequently exhibit robustness deficiencies, where slight perturbations in input (prompt+example) may lead to erroneous outputs. Current robustness testing methods face two main limitations: (1) low testing effectiveness, limiting the applicability of LLM-based software in safety-critical scenarios, and (2) insufficient naturalness of test cases, reducing the practical value of testing outcomes. To address these issues, this paper proposes ABFS, a straightforward yet effective automated testing method that, for the first time, treats the input prompts and examples as a unified whole for robustness testing. Specifically, ABFS formulates the testing process as a combinatorial optimization problem, employing Best-First Search to identify successful test cases within the perturbation space and designing a novel Adaptive control strategy to enhance test case naturalness. We evaluate the robustness testing performance of ABFS on three datasets across five threat models. On Llama2-13b, the traditional StressTest achieves only a 13.273% success rate, while ABFS attains a success rate of 98.064%, supporting a more comprehensive robustness assessment before software deployment. Compared to baseline methods, ABFS introduces fewer modifications to the original input and consistently generates test cases with superior naturalness. Furthermore, test cases generated by ABFS exhibit stronger transferability and higher testing efficiency, significantly reducing testing costs.
Large Language Models (LLMs) have reshaped code generation by synergizing their exceptional comprehension of natural language and programming syntax, thereby substantially boosting developer productivity. These advancements have prompted numerous efforts to quantitatively evaluate their coding capabilities. However, persistent challenges, such as benchmark leakage, data dissipation, and limited system accessibility, continue to impede a timely and accurate assessment. To address these limitations, we introduce CodeArena, an online evaluation framework tailored for LLM code generation. The key innovation is a collective evaluation mechanism, which dynamically recalibrates individual model scores based on the holistic performance of all participating models, mitigating score biases caused by widespread benchmark leakage. In addition, CodeArena ensures open access to all submitted solutions and test cases and provides automation-friendly APIs to streamline the code evaluation workflow. Our main contributions are: (1) a collective evaluation system for unbiased assessment, (2) a public repository of solutions and test cases, and (3) automation-ready APIs for seamless integration.
Multi-terrain cost-efficient path planning is a crucial task in robot navigation, requiring the identification of a path from the start to the goal that not only avoids obstacles but also minimizes travel costs. This is especially crucial for real-world applications where robots need to navigate diverse terrains in outdoor environments, where recharging or refueling is difficult. However, there is very limited research on this topic. In this paper, we develop a prompt-based approach, LLM-Advisor, which leverages large language models (LLMs) as effective advisors for path planning. The LLM-Advisor selectively provides suggestions, demonstrating its ability to recognize when no modifications are necessary. When suggestions are made, 70.59% of the paths suggested for the A* algorithm, 69.47% for the RRT* algorithm, and 78.70% for the LLM-A* algorithm achieve greater cost efficiency. Since LLM-Advisor may occasionally lack common sense in their suggestions, we propose two hallucination-mitigation strategies. Furthermore, we experimentally verified that GPT-4o performs poorly in zero-shot path planning, even when terrain descriptions are clearly provided, demonstrating its low spatial awareness. We also experimentally demonstrate that using an LLM as an advisor is more effective than directly integrating it into the path-planning loop. Since LLMs may generate hallucinations, using LLMs in the loop of a search-based method (such as A*) may lead to a higher number of failed paths, demonstrating that our proposed LLM-Advisor is a better choice.
Large Language Models (LLMs) have shown significant promise across various natural language processing tasks. However, their application in the field of pathology, particularly for extracting meaningful insights from unstructured medical texts such as pathology reports, remains underexplored and not well quantified. In this project, we leverage state-of-the-art language models, including the GPT family, Mistral models, and the open-source Llama models, to evaluate their performance in comprehensively analyzing pathology reports. Specifically, we assess their performance in cancer type identification, AJCC stage determination, and prognosis assessment, encompassing both information extraction and higher-order reasoning tasks. Based on a detailed analysis of their performance metrics in a zero-shot setting, we developed two instruction-tuned models: Path-llama3.1-8B and Path-GPT-4o-mini-FT. These models demonstrated superior performance in zero-shot cancer type identification, staging, and prognosis assessment compared to the other models evaluated.
Large Language Models (LLMs) have shown remarkable capabilities as autonomous agents, yet existing benchmarks either focus on single-agent tasks or are confined to narrow domains, failing to capture the dynamics of multi-agent coordination and competition. In this paper, we introduce MultiAgentBench, a comprehensive benchmark designed to evaluate LLM-based multi-agent systems across diverse, interactive scenarios. Our framework measures not only task completion but also the quality of collaboration and competition using novel, milestone-based key performance indicators. Moreover, we evaluate various coordination protocols (including star, chain, tree, and graph topologies) and innovative strategies such as group discussion and cognitive planning. Notably, gpt-4o-mini reaches the average highest task score, graph structure performs the best among coordination protocols in the research scenario, and cognitive planning improves milestone achievement rates by 3%. Code and datasets are public available at https://github.com/MultiagentBench/MARBLE.
Large language models (LLMs) excel in high-resource languages but struggle with low-resource languages (LRLs), particularly those spoken by minority communities in China, such as Tibetan, Uyghur, Kazakh, and Mongolian. To systematically track the progress in these languages, we introduce MiLiC-Eval, a benchmark designed for minority languages in China, featuring 24K instances across 9 tasks. MiLiC-Eval focuses on underrepresented writing systems and provides a fine-grained assessment of linguistic and problem-solving skills. Our evaluation reveals that LLMs perform poorly on syntax-intensive tasks and multi-script languages. We further demonstrate how MiLiC-Eval can help advance LRL research in handling diverse writing systems and understanding the process of language adaptation.
Chain-of-thought prompting has emerged as a powerful technique for enabling large language models (LLMs) to solve complex reasoning tasks. However, these reasoning chains can be verbose, raising concerns about efficiency. In response, recent works have sought to decrease response lengths through simple prompting strategies (e.g. 'be concise'). In this work, we conduct the first systematic study of the relationship between reasoning length and model performance across a diverse range of compression instructions (e.g. 'use 10 words or less' or 'remove all punctuation'). In doing so, we discover a universal tradeoff between reasoning length and accuracy that persists across even very distinct reasoning chains. We demonstrate that this tradeoff emerges from a sharp threshold behavior at the question level: each task has an intrinsic 'token complexity' - a minimal number of tokens required for successful problem-solving. We show how token complexity enables us to compute information-theoretic limits on the accuracy-compression tradeoff, and find that prompt-based compression strategies operate far from these theoretical limits. This suggests there may be significant room for improvement and our framework provides a benchmark to help researchers evaluate progress in reasoning efficiency. Our work also highlights the importance of adaptive compression -- giving shorter responses for easier questions -- and we show that token complexity is a useful tool for measuring this capability.
This paper presents an extensive examination of Parameter-Efficient Fine-Tuning (PEFT) for embedding domain specific facts into Large Language Models (LLMs), focusing on improving the fine-tuning process by categorizing question-answer (QA) pairs into Factual and Conceptual classes using a BERT-based classifier. Two distinct Llama-2 models are fine-tuned based on these classifications and evaluated using larger models like GPT-3.5 Turbo and Gemini. Our results indicate that models trained on conceptual datasets outperform those trained on factual datasets. Additionally, we compare the efficiency of two synthetic fine-tuning dataset generation techniques, D-RAG and D-Naive, with D-Naive demonstrating superior performance. Although PEFT has shown effectiveness, our research indicates that it may not be the most optimal method for embedding facts into LLMs. However, it has demonstrated exceptional performance in instruction-based tasks. Our findings are reinforced by a 1000-sample dataset in the data center domain, where the fine-tuned Llama-2 7B model significantly outperforms the baseline model in generating product recommendations. Our study highlights the importance of QA pair categorization and synthetic dataset generation techniques in enhancing the performance of LLMs in specific domains.
Knowledge editing aims to update outdated information in Large Language Models (LLMs). A representative line of study is locate-then-edit methods, which typically employ causal tracing to identify the modules responsible for recalling factual knowledge about entities. However, we find these methods are often sensitive only to changes in the subject entity, leaving them less effective at adapting to changes in relations. This limitation results in poor editing locality, which can lead to the persistence of irrelevant or inaccurate facts, ultimately compromising the reliability of LLMs. We believe this issue arises from the insufficient precision of knowledge localization. To address this, we propose a Fine-grained Neuron-level Knowledge Editing (FiNE) method that enhances editing locality without affecting overall success rates. By precisely identifying and modifying specific neurons within feed-forward networks, FiNE significantly improves knowledge localization and editing. Quantitative experiments demonstrate that FiNE efficiently achieves better overall performance compared to existing techniques, providing new insights into the localization and modification of knowledge within LLMs.
Large language models (LLMs) can answer questions and reason about complex tasks, also from the scientific domain. We assess several multimodal LLMs (MLLMs) on ScienceQA and find that Gemini models show the highest accuracy with little context, and the highest textual similarity to human explanations with richer context. Adapter-tuning of smaller MLLMs did not lead to any reliable performance. Training from Gemini outputs consistently underperformed training from the original data.
Recent developments in computer graphics, machine learning, and sensor technologies enable numerous opportunities for extended reality (XR) setups for everyday life, from skills training to entertainment. With large corporations offering affordable consumer-grade head-mounted displays (HMDs), XR will likely become pervasive, and HMDs will develop as personal devices like smartphones and tablets. However, having intelligent spaces and naturalistic interactions in XR is as important as technological advances so that users grow their engagement in virtual and augmented spaces. To this end, large language model (LLM)--powered non-player characters (NPCs) with speech-to-text (STT) and text-to-speech (TTS) models bring significant advantages over conventional or pre-scripted NPCs for facilitating more natural conversational user interfaces (CUIs) in XR. This paper provides the community with an open-source, customizable, extendable, and privacy-aware Unity package, CUIfy, that facilitates speech-based NPC-user interaction with widely used LLMs, STT, and TTS models. Our package also supports multiple LLM-powered NPCs per environment and minimizes latency between different computational models through streaming to achieve usable interactions between users and NPCs. We publish our source code in the following repository: https://gitlab.lrz.de/hctl/cuify
Vision-and-Language Navigation (VLN) is an essential skill for embodied agents, allowing them to navigate in 3D environments following natural language instructions. High-performance navigation models require a large amount of training data, the high cost of manually annotating data has seriously hindered this field. Therefore, some previous methods translate trajectory videos into step-by-step instructions for expanding data, but such instructions do not match well with users' communication styles that briefly describe destinations or state specific needs. Moreover, local navigation trajectories overlook global context and high-level task planning. To address these issues, we propose NavRAG, a retrieval-augmented generation (RAG) framework that generates user demand instructions for VLN. NavRAG leverages LLM to build a hierarchical scene description tree for 3D scene understanding from global layout to local details, then simulates various user roles with specific demands to retrieve from the scene tree, generating diverse instructions with LLM. We annotate over 2 million navigation instructions across 861 scenes and evaluate the data quality and navigation performance of trained models.
Video Temporal Grounding (VTG) is a crucial capability for video understanding models and plays a vital role in downstream tasks such as video browsing and editing. To effectively handle various tasks simultaneously and enable zero-shot prediction, there is a growing trend in employing video LLMs for VTG tasks. However, current video LLM-based methods rely exclusively on natural language generation, lacking the ability to model the clear structure inherent in videos, which restricts their effectiveness in tackling VTG tasks. To address this issue, this paper first formally introduces causal event modeling framework, which represents video LLM outputs as sequences of events, and predict the current event using previous events, video inputs, and textural instructions. Each event consists of three components: timestamps, salient scores, and textual captions. We then propose a novel task-interleaved video LLM called TRACE to effectively implement the causal event modeling framework in practice. The TRACE process visual frames, timestamps, salient scores, and text as distinct tasks, employing various encoders and decoding heads for each. Task tokens are arranged in an interleaved sequence according to the causal event modeling framework's formulation. Extensive experiments on various VTG tasks and datasets demonstrate the superior performance of TRACE compared to state-of-the-art video LLMs. Our model and code are available at https://github.com/gyxxyg/TRACE.
Large language models (LLMs) are evolving into autonomous decision-makers, raising concerns about catastrophic risks in high-stakes scenarios, particularly in Chemical, Biological, Radiological and Nuclear (CBRN) domains. Based on the insight that such risks can originate from trade-offs between the agent's Helpful, Harmlessness and Honest (HHH) goals, we build a novel three-stage evaluation framework, which is carefully constructed to effectively and naturally expose such risks. We conduct 14,400 agentic simulations across 12 advanced LLMs, with extensive experiments and analysis. Results reveal that LLM agents can autonomously engage in catastrophic behaviors and deception, without being deliberately induced. Furthermore, stronger reasoning abilities often increase, rather than mitigate, these risks. We also show that these agents can violate instructions and superior commands. On the whole, we empirically prove the existence of catastrophic risks in autonomous LLM agents. We will release our code upon request.
Emotion recognition in conversation (ERC), the task of discerning human emotions for each utterance within a conversation, has garnered significant attention in human-computer interaction systems. Previous ERC studies focus on speaker-specific information that predominantly stems from relationships among utterances, which lacks sufficient information around conversations. Recent research in ERC has sought to exploit pre-trained large language models (LLMs) with speaker modelling to comprehend emotional states. Although these methods have achieved encouraging results, the extracted speaker-specific information struggles to indicate emotional dynamics. In this paper, motivated by the fact that speaker characteristics play a crucial role and LLMs have rich world knowledge, we present LaERC-S, a novel framework that stimulates LLMs to explore speaker characteristics involving the mental state and behavior of interlocutors, for accurate emotion predictions. To endow LLMs with this knowledge information, we adopt the two-stage learning to make the models reason speaker characteristics and track the emotion of the speaker in complex conversation scenarios. Extensive experiments on three benchmark datasets demonstrate the superiority of LaERC-S, reaching the new state-of-the-art.
Large Language Models (LLMs), trained on a large amount of corpus, have demonstrated remarkable abilities. However, it may not be sufficient to directly apply open-source LLMs like Llama to certain real-world scenarios, since most of them are trained for \emph{general} purposes. Thus, the demands for customizing publicly available LLMs emerge, but are currently under-studied. In this work, we consider customizing pre-trained LLMs with new human preferences. Specifically, the LLM should not only meet the new preference but also preserve its original capabilities after customization. Drawing inspiration from the observation that human preference can be expressed as a reward model, we propose to cast LLM customization as optimizing the sum of two reward functions, one of which (denoted as $r_1$) was used to pre-train the LLM while the other (denoted as $r_2$) characterizes the new human preference. The obstacle here is that both reward functions are unknown, making the application of modern reinforcement learning methods infeasible. Thanks to the residual Q-learning framework, we can restore the customized LLM with the pre-trained LLM and the \emph{residual Q-function} without the reward function $r_1$. Moreover, we find that for a fixed pre-trained LLM, the reward function $r_2$ can be derived from the residual Q-function, enabling us to directly learn the residual Q-function from the new human preference data upon the Bradley-Terry model. We name our method Q-Adapter as it introduces an adapter module to approximate the residual Q-function for customizing the pre-trained LLM towards the new preference. Experiments based on the Llama-3.1 model on the DSP dataset and HH-RLHF dataset illustrate the superior effectiveness of Q-Adapter on both retaining existing knowledge and learning new preferences. Code is available at https://github.com/mansicer/Q-Adapter.
Large Language Models (LLMs) have excelled in various tasks but are still vulnerable to jailbreaking attacks, where attackers create jailbreak prompts to mislead the model to produce harmful or offensive content. Current jailbreak methods either rely heavily on manually crafted templates, which pose challenges in scalability and adaptability, or struggle to generate semantically coherent prompts, making them easy to detect. Additionally, most existing approaches involve lengthy prompts, leading to higher query costs. In this paper, to remedy these challenges, we introduce a novel jailbreaking attack framework called PAPILLON, which is an automated, black-box jailbreaking attack framework that adapts the black-box fuzz testing approach with a series of customized designs. Instead of relying on manually crafted templates,PAPILLON starts with an empty seed pool, removing the need to search for any related jailbreaking templates. We also develop three novel question-dependent mutation strategies using an LLM helper to generate prompts that maintain semantic coherence while significantly reducing their length. Additionally, we implement a two-level judge module to accurately detect genuine successful jailbreaks. We evaluated PAPILLON on 7 representative LLMs and compared it with 5 state-of-the-art jailbreaking attack strategies. For proprietary LLM APIs, such as GPT-3.5 turbo, GPT-4, and Gemini-Pro, PAPILLONs achieves attack success rates of over 90%, 80%, and 74%, respectively, exceeding existing baselines by more than 60\%. Additionally, PAPILLON can maintain high semantic coherence while significantly reducing the length of jailbreak prompts. When targeting GPT-4, PAPILLON can achieve over 78% attack success rate even with 100 tokens. Moreover, PAPILLON demonstrates transferability and is robust to state-of-the-art defenses. Code: https://github.com/aaFrostnova/Papillon
As users increasingly seek guidance from LLMs for decision-making in daily life, many of these decisions are not clear-cut and depend significantly on the personal values and ethical standards of people. We present DailyDilemmas, a dataset of 1,360 moral dilemmas encountered in everyday life. Each dilemma presents two possible actions, along with affected parties and relevant human values for each action. Based on these dilemmas, we gather a repository of human values covering diverse everyday topics, such as interpersonal relationships, workplace, and environmental issues. With DailyDilemmas, we evaluate LLMs on these dilemmas to determine what action they will choose and the values represented by these action choices. Then, we analyze values through the lens of five theoretical frameworks inspired by sociology, psychology, and philosophy, including the World Values Survey, Moral Foundations Theory, Maslow's Hierarchy of Needs, Aristotle's Virtues, and Plutchik's Wheel of Emotions. For instance, we find LLMs are most aligned with self-expression over survival in World Values Survey and care over loyalty in Moral Foundations Theory. Interestingly, we find substantial preference differences in models for some core values. For example, for truthfulness, Mixtral-8x7B neglects it by 9.7% while GPT-4-turbo selects it by 9.4%. We also study the recent guidance released by OpenAI (ModelSpec), and Anthropic (Constitutional AI) to understand how their designated principles reflect their models' actual value prioritization when facing nuanced moral reasoning in daily-life settings. Finally, we find that end users cannot effectively steer such prioritization using system prompts.
While LLMs have exhibited strong performance on various NLP tasks, it is noteworthy that most of these tasks rely on utilizing the vast amount of knowledge encoded in LLMs' parameters, rather than solving new problems without prior knowledge. In cognitive research, the latter ability is referred to as fluid intelligence, which is considered to be critical for assessing human intelligence. Recent research on fluid intelligence assessments has highlighted significant deficiencies in LLMs' abilities. In this paper, we analyze the challenges LLMs face in demonstrating fluid intelligence through controlled experiments, using the most representative ARC task as an example. Our study revealed three major limitations in existing LLMs: limited ability for skill composition, unfamiliarity with abstract input formats, and the intrinsic deficiency of left-to-right decoding. Our data and code can be found in https://wujunjie1998.github.io/araoc-benchmark.github.io/.
The rapid advancement of Large Language Models (LLMs) has driven growing demand for processing extended context sequences in contemporary applications. However, this progress faces two major challenges: performance degradation due to sequence lengths out-of-distribution, and excessively long inference times caused by the quadratic computational complexity of attention. These issues hinder the application of LLMs in long-context scenarios. In this paper, we propose Dynamic Token-Level KV Cache Selection (TokenSelect), a training-free method for efficient and accurate long-context inference. TokenSelect builds upon the observation of non-contiguous attention sparsity, using Query-Key dot products to measure per-head KV Cache criticality at token-level. By per-head soft voting mechanism, TokenSelect selectively involves a few critical KV cache tokens in attention calculation without sacrificing accuracy. To further accelerate TokenSelect, we design the Selection Cache based on observations of consecutive Query similarity and implemented efficient dot product kernel, significantly reducing the overhead. A comprehensive evaluation of TokenSelect demonstrates up to 23.84x speedup in attention computation and up to 2.28x acceleration in end-to-end latency, while providing superior performance compared to state-of-the-art long-context inference methods.
While large language model (LLM) agents can effectively use external tools for complex real-world tasks, they require memory systems to leverage historical experiences. Current memory systems enable basic storage and retrieval but lack sophisticated memory organization, despite recent attempts to incorporate graph databases. Moreover, these systems' fixed operations and structures limit their adaptability across diverse tasks. To address this limitation, this paper proposes a novel agentic memory system for LLM agents that can dynamically organize memories in an agentic way. Following the basic principles of the Zettelkasten method, we designed our memory system to create interconnected knowledge networks through dynamic indexing and linking. When a new memory is added, we generate a comprehensive note containing multiple structured attributes, including contextual descriptions, keywords, and tags. The system then analyzes historical memories to identify relevant connections, establishing links where meaningful similarities exist. Additionally, this process enables memory evolution - as new memories are integrated, they can trigger updates to the contextual representations and attributes of existing historical memories, allowing the memory network to continuously refine its understanding. Our approach combines the structured organization principles of Zettelkasten with the flexibility of agent-driven decision making, allowing for more adaptive and context-aware memory management. Empirical experiments on six foundation models show superior improvement against existing SOTA baselines. The source code is available at https://github.com/WujiangXu/AgenticMemory.
Reinforcement Learning with Human Feedback (RLHF) has achieved great success in aligning large language models (LLMs) with human preferences. Prevalent RLHF approaches are reward-based, following the Bradley-Terry (BT) model assumption, which may not fully capture the complexity of human preferences. In this paper, we explore RLHF under a general preference framework and approach it from a game-theoretic perspective. Specifically, we formulate the problem as a two-player game and propose a novel online algorithm, iterative Nash policy optimization (INPO). The key idea is to let the policy play against itself via no-regret learning, thereby approximating the Nash policy. Unlike previous methods, INPO bypasses the need for estimating the expected win rate for individual responses, which typically incurs high computational or annotation costs. Instead, we introduce a new loss objective that is directly minimized over a preference dataset. We provide theoretical analysis for our approach and demonstrate its effectiveness through experiments on various representative benchmarks. With an LLaMA-3-8B-based SFT model, INPO achieves a 42.6% length-controlled win rate on AlpacaEval 2.0 and a 37.8% win rate on Arena-Hard, showing substantial improvement over the state-of-the-art online RLHF algorithms.
LLM-as-a-Judge uses a large language model (LLM) to select the best response from a set of candidates for a given question. LLM-as-a-Judge has many applications such as LLM-powered search, reinforcement learning with AI feedback (RLAIF), and tool selection. In this work, we propose JudgeDeceiver, an optimization-based prompt injection attack to LLM-as-a-Judge. JudgeDeceiver injects a carefully crafted sequence into an attacker-controlled candidate response such that LLM-as-a-Judge selects the candidate response for an attacker-chosen question no matter what other candidate responses are. Specifically, we formulate finding such sequence as an optimization problem and propose a gradient based method to approximately solve it. Our extensive evaluation shows that JudgeDeceive is highly effective, and is much more effective than existing prompt injection attacks that manually craft the injected sequences and jailbreak attacks when extended to our problem. We also show the effectiveness of JudgeDeceiver in three case studies, i.e., LLM-powered search, RLAIF, and tool selection. Moreover, we consider defenses including known-answer detection, perplexity detection, and perplexity windowed detection. Our results show these defenses are insufficient, highlighting the urgent need for developing new defense strategies. Our implementation is available at this repository: https://github.com/ShiJiawenwen/JudgeDeceiver.
Large Language Models (LLMs) have exhibited significant potential in performing diverse tasks, including the ability to call functions or use external tools to enhance their performance. While current research on function calling by LLMs primarily focuses on single-turn interactions, this paper addresses the overlooked necessity for LLMs to engage in multi-turn function calling--critical for handling compositional, real-world queries that require planning with functions but not only use functions. To facilitate this, we introduce an approach, BUTTON, which generates synthetic compositional instruction tuning data via bottom-up instruction construction and top-down trajectory generation. In the bottom-up phase, we generate simple atomic tasks based on real-world scenarios and build compositional tasks using heuristic strategies based on atomic tasks. Corresponding function definitions are then synthesized for these compositional tasks. The top-down phase features a multi-agent environment where interactions among simulated humans, assistants, and tools are utilized to gather multi-turn function calling trajectories. This approach ensures task compositionality and allows for effective function and trajectory generation by examining atomic tasks within compositional tasks. We produce a dataset BUTTONInstruct comprising 8k data points and demonstrate its effectiveness through extensive experiments across various LLMs.
This paper presents a comprehensive performance evaluation of Large Language Models (LLMs) in solving programming challenges from Leetcode, a widely used platform for algorithm practice and technical interviews. We began by crawling the Leetcode website to collect a diverse set of problems encompassing various difficulty levels and topics. Using this dataset, we generated solutions with multiple LLMs, including GPT-4 and GPT-3.5-turbo (ChatGPT-turbo). The generated solutions were systematically evaluated for correctness and efficiency. We employed the pass@k metric to assess the success rates within a given number of attempts and analyzed the runtime performance of the solutions. Our results highlight the strengths and limitations of current LLMs [10] in code generation and problem-solving tasks, providing insights into their potential applications and areas for improvement in automated programming assistance.
Natural language is composed of words, but modern large language models (LLMs) process sub-words as input. A natural question raised by this discrepancy is whether LLMs encode words internally, and if so how. We present evidence that LLMs engage in an intrinsic detokenization process, where sub-word sequences are combined into coherent whole-word representations at their last token. Our experiments show that this process primarily takes place within the early and middle layers of the model. We further demonstrate its robustness to arbitrary splits (e.g., "cats" to "ca" and "ts"), typos, and importantly-to out-of-vocabulary words: when feeding the last token internal representations of such words to the model as input, it can "understand" them as the complete word despite never seeing such representations as input during training. Our findings suggest that LLMs maintain a latent vocabulary beyond the tokenizer's scope. These insights provide a practical, finetuning-free application for expanding the vocabulary of pre-trained models. By enabling the addition of new vocabulary words, we reduce input length and inference iterations, which reduces both space and model latency, with little to no loss in model accuracy.
Despite recent breakthroughs in reasoning-enhanced large language models (LLMs) like DeepSeek-R1, incorporating inference-time reasoning into machine translation (MT), where human translators naturally employ structured, multi-layered reasoning chain-of-thoughts (CoTs), is yet underexplored. Existing methods either design a fixed CoT tailored for a specific MT sub-task (e.g., literature translation), or rely on synthesizing CoTs unaligned with humans, limiting their adaptability to diverse translation scenarios. This paper introduces R1-Translator (R1-T1), a novel framework to achieve inference-time reasoning for general MT via reinforcement learning (RL) with human-aligned CoTs comprising six common patterns. Our approach pioneers three innovations: (1) extending reasoning-based translation beyond MT sub-tasks to six languages and diverse tasks (e.g., legal/medical domain adaptation, idiom resolution); (2) formalizing six expert-curated CoT templates that mirror hybrid human strategies like context-aware paraphrasing and back translation; and (3) enabling self-evolving CoT discovery through RL. Experimental results indicate a steady translation performance improvement in 11 languages and 40 translation directions on Flores-101 test set, especially on the languages unseen from training.
Large language models (LLMs) have shown remarkable improvements in reasoning and many existing benchmarks have been addressed by models such as o1 and o3 either fully or partially. However, a majority of these benchmarks emphasize deductive reasoning, including mathematical and coding tasks in which rules such as mathematical axioms or programming syntax are clearly defined, based on which LLMs can plan and apply these rules to arrive at a solution. In contrast, inductive reasoning, where one infers the underlying rules from observed data, remains less explored. Such inductive processes lie at the heart of scientific discovery, as they enable researchers to extract general principles from empirical observations. To assess whether LLMs possess this capacity, we introduce InductionBench, a new benchmark designed to evaluate the inductive reasoning ability of LLMs. Our experimental findings reveal that even the most advanced models available struggle to master the simplest complexity classes within the subregular hierarchy of functions, highlighting a notable deficiency in current LLMs' inductive reasoning capabilities. Coda and data are available https://github.com/Wenyueh/inductive_reasoning_benchmark.
Reading assessments are essential for enhancing students' comprehension, yet many EdTech applications focus mainly on outcome-based metrics, providing limited insights into student behavior and cognition. This study investigates the use of multimodal data sources -- including eye-tracking data, learning outcomes, assessment content, and teaching standards -- to derive meaningful reading insights. We employ unsupervised learning techniques to identify distinct reading behavior patterns, and then a large language model (LLM) synthesizes the derived information into actionable reports for educators, streamlining the interpretation process. LLM experts and human educators evaluate these reports for clarity, accuracy, relevance, and pedagogical usefulness. Our findings indicate that LLMs can effectively function as educational analysts, turning diverse data into teacher-friendly insights that are well-received by educators. While promising for automating insight generation, human oversight remains crucial to ensure reliability and fairness. This research advances human-centered AI in education, connecting data-driven analytics with practical classroom applications.
Pro-environmental behavior (PEB) is vital to combat climate change, yet turning awareness into intention and action remains elusive. We explore large language models (LLMs) as tools to promote PEB, comparing their impact across 3,200 participants: real humans (n=1,200), simulated humans based on actual participant data (n=1,200), and fully synthetic personas (n=1,200). All three participant groups faced personalized or standard chatbots, or static statements, employing four persuasion strategies (moral foundations, future self-continuity, action orientation, or "freestyle" chosen by the LLM). Results reveal a "synthetic persuasion paradox": synthetic and simulated agents significantly affect their post-intervention PEB stance, while human responses barely shift. Simulated participants better approximate human trends but still overestimate effects. This disconnect underscores LLM's potential for pre-evaluating PEB interventions but warns of its limits in predicting real-world behavior. We call for refined synthetic modeling and sustained and extended human trials to align conversational AI's promise with tangible sustainability outcomes.
Existing challenges in misinformation exposure and susceptibility vary across demographic groups, as some populations are more vulnerable to misinformation than others. Large language models (LLMs) introduce new dimensions to these challenges through their ability to generate persuasive content at scale and reinforcing existing biases. This study investigates the bidirectional persuasion dynamics between LLMs and humans when exposed to misinformative content. We analyze human-to-LLM influence using human-stance datasets and assess LLM-to-human influence by generating LLM-based persuasive arguments. Additionally, we use a multi-agent LLM framework to analyze the spread of misinformation under persuasion among demographic-oriented LLM agents. Our findings show that demographic factors influence susceptibility to misinformation in LLMs, closely reflecting the demographic-based patterns seen in human susceptibility. We also find that, similar to human demographic groups, multi-agent LLMs exhibit echo chamber behavior. This research explores the interplay between humans and LLMs, highlighting demographic differences in the context of misinformation and offering insights for future interventions.
Can LLMs pick up language structure from examples? Evidence in prior work seems to indicate yes, as pretrained models repeatedly demonstrate the ability to adapt to new language structures and vocabularies. However, this line of research typically considers languages that are present within common pretraining datasets, or otherwise share notable similarities with these seen languages. In contrast, in this work we attempt to measure models' language understanding capacity while circumventing the risk of dataset recall. We parameterize large families of language tasks recognized by deterministic finite automata (DFAs), and can thus sample novel language reasoning problems to fairly evaulate LLMs regardless of training data. We find that, even in the strikingly simple setting of 3-state DFAs, LLMs underperform unparameterized ngram models on both language recognition and synthesis tasks. These results suggest that LLMs struggle to match the ability of basic language models in recognizing and reasoning over languages that are sufficiently distinct from the ones they see at training time, underscoring the distinction between learning individual languages and possessing a general theory of language.
Large Language Models (LLMs) have made significant strides in natural language processing, and a precise understanding of the internal mechanisms driving their success is essential. In this work, we analyze the trajectories of token embeddings as they pass through transformer blocks, linearizing the system along these trajectories through their Jacobian matrices. By examining the relationships between these block Jacobians, we uncover the phenomenon of \textbf{transformer block coupling} in a multitude of LLMs, characterized by the coupling of their top singular vectors across tokens and depth. Our findings reveal that coupling \textit{positively correlates} with model performance, and that this relationship is stronger than with other hyperparameters such as parameter count, model depth, and embedding dimension. We further investigate the emergence of these properties through training, observing the progressive development of coupling, as well as increased linearity and layer-wise exponential growth in the token trajectories. Additionally, experiments with ViTs further validate emergence of coupling and its correlation between coupling and generalization, complementing our findings in LLMs. Collectively, these insights provide a novel perspective on token interactions in transformers and open directions for studying and improving training and generalization.
Social biases and belief-driven behaviors can significantly impact Large Language Models (LLMs) decisions on several tasks. As LLMs are increasingly used in multi-agent systems for societal simulations, their ability to model fundamental group psychological characteristics remains critical yet under-explored. In this study, we present a multi-agent framework that simulates belief congruence, a classical group psychology theory that plays a crucial role in shaping societal interactions and preferences. Our findings reveal that LLMs exhibit amplified belief congruence compared to humans, across diverse contexts. We further investigate the implications of this behavior on two downstream tasks: (1) misinformation dissemination and (2) LLM learning, finding that belief congruence in LLMs increases misinformation dissemination and impedes learning. To mitigate these negative impacts, we propose strategies inspired by: (1) contact hypothesis, (2) accuracy nudges, and (3) global citizenship framework. Our results show that the best strategies reduce misinformation dissemination by up to 37% and enhance learning by 11%. Bridging social psychology and AI, our work provides insights to navigate real-world interactions using LLMs while addressing belief-driven biases.
Large language models (LLMs) demonstrate remarkable emergent abilities to perform in-context learning across various tasks, including time series forecasting. This work investigates LLMs' ability to estimate probability density functions (PDFs) from data observed in-context; such density estimation (DE) is a fundamental task underlying many probabilistic modeling problems. We leverage the Intensive Principal Component Analysis (InPCA) to visualize and analyze the in-context learning dynamics of LLaMA-2 models. Our main finding is that these LLMs all follow similar learning trajectories in a low-dimensional InPCA space, which are distinct from those of traditional density estimation methods like histograms and Gaussian kernel density estimation (KDE). We interpret the LLaMA in-context DE process as a KDE with an adaptive kernel width and shape. This custom kernel model captures a significant portion of LLaMA's behavior despite having only two parameters. We further speculate on why LLaMA's kernel width and shape differs from classical algorithms, providing insights into the mechanism of in-context probabilistic reasoning in LLMs. Our codebase, along with a 3D visualization of an LLM's in-context learning trajectory, is publicly available at https://github.com/AntonioLiu97/LLMICL_inPCA
Distilling advanced Large Language Models' instruction-following capabilities into smaller models using a selected subset has become a mainstream approach in model training. While existing synthetic instruction data selection strategies rely mainly on single-dimensional signals (i.e., reward scores, model perplexity), they fail to capture the complexity of instruction-following across diverse fields. Therefore, we investigate more diverse signals to capture comprehensive instruction-response pair characteristics and propose three foundational metrics that leverage Multi-LLM wisdom, informed by (1) diverse LLM responses and (2) reward model assessment. Building upon base metrics, we propose CrowdSelect, an integrated metric incorporating a clustering-based approach to maintain response diversity. Our comprehensive experiments demonstrate that our foundation metrics consistently improve performance across 4 base models on MT-bench and Arena-Hard. CrowdSelect, efficiently incorporating all metrics, achieves state-of-the-art performance in both Full and LoRA fine-tuning, showing improvements of 4.81% on Arena-Hard and 11.1% on MT-bench with Llama-3.2-3b-instruct. We hope our findings will bring valuable insights for future research in this direction. Code are available at https://github.com/listentm/crowdselect.
Large language models (LLMs) exhibit remarkable similarity to neural activity in the human language network. However, the key properties of language shaping brain-like representations, and their evolution during training as a function of different tasks remain unclear. We here benchmark 34 training checkpoints spanning 300B tokens across 8 different model sizes to analyze how brain alignment relates to linguistic competence. Specifically, we find that brain alignment tracks the development of formal linguistic competence -- i.e., knowledge of linguistic rules -- more closely than functional linguistic competence. While functional competence, which involves world knowledge and reasoning, continues to develop throughout training, its relationship with brain alignment is weaker, suggesting that the human language network primarily encodes formal linguistic structure rather than broader cognitive functions. We further show that model size is not a reliable predictor of brain alignment when controlling for feature size and find that the correlation between next-word prediction, behavioral alignment and brain alignment fades once models surpass human language proficiency. Finally, using the largest set of rigorous neural language benchmarks to date, we show that language brain alignment benchmarks remain unsaturated, highlighting opportunities for improving future models. Taken together, our findings suggest that the human language network is best modeled by formal, rather than functional, aspects of language.
We formalize a new concept for LLMs, context-enhanced learning. It involves standard gradient-based learning on text except that the context is enhanced with additional data on which no auto-regressive gradients are computed. This setting is a gradient-based analog of usual in-context learning (ICL) and appears in some recent works. Using a multi-step reasoning task, we prove in a simplified setting that context-enhanced learning can be exponentially more sample-efficient than standard learning when the model is capable of ICL. At a mechanistic level, we find that the benefit of context-enhancement arises from a more accurate gradient learning signal. We also experimentally demonstrate that it appears hard to detect or recover learning materials that were used in the context during training. This may have implications for data security as well as copyright.
Alzheimer's Disease (AD) has become an increasingly critical global health concern, which necessitates effective monitoring solutions in smart health applications. However, the development of such solutions is significantly hindered by the scarcity of AD-specific activity datasets. To address this challenge, we propose SHADE-AD, a Large Language Model (LLM) framework for Synthesizing Human Activity Datasets Embedded with AD features. Leveraging both public datasets and our own collected data from 99 AD patients, SHADE-AD synthesizes human activity videos that specifically represent AD-related behaviors. By employing a three-stage training mechanism, it broadens the range of activities beyond those collected from limited deployment settings. We conducted comprehensive evaluations of the generated dataset, demonstrating significant improvements in downstream tasks such as Human Activity Recognition (HAR) detection, with enhancements of up to 79.69%. Detailed motion metrics between real and synthetic data show strong alignment, validating the realism and utility of the synthesized dataset. These results underscore SHADE-AD's potential to advance smart health applications by providing a cost-effective, privacy-preserving solution for AD monitoring.
Fuzz testing is a crucial component of software security assessment, yet its effectiveness heavily relies on valid fuzz drivers and diverse seed inputs. Recent advancements in Large Language Models (LLMs) offer transformative potential for automating fuzz testing (LLM4Fuzz), particularly in generating drivers and seeds. However, current LLM4Fuzz solutions face critical reliability challenges, including low driver validity rates and seed quality trade-offs, hindering their practical adoption. This paper aims to examine the reliability bottlenecks of LLM-driven fuzzing and explores potential research directions to address these limitations. It begins with an overview of the current development of LLM4SE and emphasizes the necessity for developing reliable LLM4Fuzz solutions. Following this, the paper envisions a vision where reliable LLM4Fuzz transforms the landscape of software testing and security for industry, software development practitioners, and economic accessibility. It then outlines a road ahead for future research, identifying key challenges and offering specific suggestions for the researchers to consider. This work strives to spark innovation in the field, positioning reliable LLM4Fuzz as a fundamental component of modern software testing.
Evaluating Personalized Tool-Augmented LLMs from the Perspectives of Personalization and Proactivity
Personalized tool utilization is essential for aligning large language models (LLMs) with user preference in interaction scenarios with various tools. However, most of the current benchmarks primarily focus on either personalization of text generation or direct tool-utilizing, without considering both. In this work, we introduce a novel benchmark ETAPP for evaluating personalized tool invocation, establishing a sandbox environment, and a comprehensive dataset of 800 testing cases covering diverse user profiles. To improve the accuracy of our evaluation, we propose a key-point-based LLM evaluation method, mitigating biases in the LLM-as-a-judge system by manually annotating key points for each test case and providing them to LLM as the reference. Additionally, we evaluate the excellent LLMs and provide an in-depth analysis. Furthermore, we investigate the impact of different tool-invoking strategies on LLMs' personalization performance and the effects of fine-tuning in our task. The effectiveness of our preference-setting and key-point-based evaluation method is also validated. Our findings offer insights into improving personalized LLM agents. Our Code is available at https://github.com/hypasd-art/ETAPP.
We are witnessing a new era where problem-solving and cognitive tasks are being increasingly delegated to Large Language Models (LLMs) across diverse domains, ranging from code generation to holiday planning. This trend also creates a demand for the ubiquitous execution of LLM-powered applications in a wide variety of environments in which traditional terrestrial 2D networking infrastructures may prove insufficient. A promising solution in this context is to extend edge computing into a 3D setting to include aerial platforms organized in multiple layers, a paradigm we refer to as air computing, to augment local devices for running LLM and Generative AI (GenAI) applications. This approach alleviates the strain on existing infrastructure while enhancing service efficiency by offloading computational tasks to the corresponding air units such as UAVs. Furthermore, the coordinated deployment of various air units can significantly improve the Quality of Experience (QoE) by ensuring seamless, adaptive, and resilient task execution. In this study, we investigate the synergy between LLM-based applications and air computing, exploring their potential across various use cases. Additionally, we present a disaster response case study demonstrating how the collaborative utilization of LLMs and air computing can significantly improve outcomes in critical situations.
The proliferation of misinformation on social media has raised significant societal concerns, necessitating robust detection mechanisms. Large Language Models such as GPT-4 and LLaMA2 have been envisioned as possible tools for detecting misinformation based on their advanced natural language understanding and reasoning capabilities. This paper conducts a comparison of LLM-based approaches to detecting misinformation between text-based, multimodal, and agentic approaches. We evaluate the effectiveness of fine-tuned models, zero-shot learning, and systematic fact-checking mechanisms in detecting misinformation across different topic domains like public health, politics, and finance. We also discuss scalability, generalizability, and explainability of the models and recognize key challenges such as hallucination, adversarial attacks on misinformation, and computational resources. Our findings point towards the importance of hybrid approaches that pair structured verification protocols with adaptive learning techniques to enhance detection accuracy and explainability. The paper closes by suggesting potential avenues of future work, including real-time tracking of misinformation, federated learning, and cross-platform detection models.
Multi-Agent Pathfinding (MAPF) is a core challenge in multi-agent systems. Existing learning-based MAPF methods often struggle with scalability, particularly when addressing complex scenarios that are prone to deadlocks. To address these challenges, we introduce LLMDR (LLM-Driven Deadlock Detection and Resolution), an approach designed to resolve deadlocks and improve the performance of learnt MAPF models. LLMDR integrates the inference capabilities of large language models (LLMs) with learnt MAPF models and prioritized planning, enabling it to detect deadlocks and provide customized resolution strategies. We evaluate LLMDR on standard MAPF benchmark maps with varying agent numbers, measuring its performance when combined with several base models. The results demonstrate that LLMDR improves the performance of learnt MAPF models, particularly in deadlock-prone scenarios, with notable improvements in success rates. These findings show the potential of integrating LLMs to improve the scalability of learning-based MAPF methods. The source code for LLMDR is available at: https://github.com/ssbacc/llmdr-dhc
Knowledge gaps often arise during communication due to diverse backgrounds, knowledge bases, and vocabularies. With recent LLM developments, providing real-time knowledge support is increasingly viable, but is challenging due to shared and individual cognitive limitations (e.g., attention, memory, and comprehension) and the difficulty in understanding the user's context and internal knowledge. To address these challenges, we explore the key question of understanding how people want to receive real-time knowledge support. We built StopGap -- a prototype that provides real-time knowledge support for explaining jargon words in videos -- to conduct a design probe study (N=24) that explored multiple visual knowledge representation formats. Our study revealed individual differences in preferred representations and highlighted the importance of user agency, personalization, and mixed-initiative assistance. Based on our findings, we map out six key design dimensions for real-time LLM knowledge support systems and offer insights for future research in this space.
Large Language Models (LLMs), such as ChatGPT, exhibit advanced capabilities in generating text, images, and videos. However, their effective use remains constrained by challenges in prompt formulation, personalization, and opaque decision-making processes. To investigate these challenges and identify design opportunities, we conducted a two-phase qualitative study. In Phase 1, we performed in-depth interviews with eight everyday LLM users after they engaged in structured tasks using ChatGPT across both familiar and unfamiliar domains. Our findings revealed key user difficulties in constructing effective prompts, iteratively refining AI-generated responses, and assessing response reliability especially in domains beyond users' expertise. Informed by these insights, we designed a high-fidelity prototype incorporating Reflective Prompting, Section Regeneration, Input-Output Mapping, Confidence Indicators, and a Customization Panel. In Phase 2, user testing of the prototype indicated that these interface-level improvements may prove useful for reducing cognitive load, increasing transparency, and fostering more intuitive and collaborative human-AI interactions. Our study contributes to the growing discourse on human-centred AI, advocating for human-LLM interactions that enhance user agency, transparency, and co-creative interaction, ultimately supporting more intuitive, accessible, and trustworthy generative AI systems.
Traditionally, digital hardware designs are written in the Verilog hardware description language (HDL) and debugged manually by engineers. This can be time-consuming and error-prone for complex designs. Large Language Models (LLMs) are emerging as a potential tool to help generate fully functioning HDL code, but most works have focused on generation in the single-shot capacity: i.e., run and evaluate, a process that does not leverage debugging and, as such, does not adequately reflect a realistic development process. In this work, we evaluate the ability of LLMs to leverage feedback from electronic design automation (EDA) tools to fix mistakes in their own generated Verilog. To accomplish this, we present an open-source, highly customizable framework, AutoChip, which combines conversational LLMs with the output from Verilog compilers and simulations to iteratively generate and repair Verilog. To determine the success of these LLMs we leverage the VerilogEval benchmark set. We evaluate four state-of-the-art conversational LLMs, focusing on readily accessible commercial models. EDA tool feedback proved to be consistently more effective than zero-shot prompting only with GPT-4o, the most computationally complex model we evaluated. In the best case, we observed a 5.8% increase in the number of successful designs with a 34.2% decrease in cost over the best zero-shot results. Mixing smaller models with this larger model at the end of the feedback iterations resulted in equally as much success as with GPT-4o using feedback, but incurred 41.9% lower cost (corresponding to an overall decrease in cost over zero-shot by 89.6%).
The reliance of popular programming languages such as Python and JavaScript on centralized package repositories and open-source software, combined with the emergence of code-generating Large Language Models (LLMs), has created a new type of threat to the software supply chain: package hallucinations. These hallucinations, which arise from fact-conflicting errors when generating code using LLMs, represent a novel form of package confusion attack that poses a critical threat to the integrity of the software supply chain. This paper conducts a rigorous and comprehensive evaluation of package hallucinations across different programming languages, settings, and parameters, exploring how a diverse set of models and configurations affect the likelihood of generating erroneous package recommendations and identifying the root causes of this phenomenon. Using 16 popular LLMs for code generation and two unique prompt datasets, we generate 576,000 code samples in two programming languages that we analyze for package hallucinations. Our findings reveal that that the average percentage of hallucinated packages is at least 5.2% for commercial models and 21.7% for open-source models, including a staggering 205,474 unique examples of hallucinated package names, further underscoring the severity and pervasiveness of this threat. To overcome this problem, we implement several hallucination mitigation strategies and show that they are able to significantly reduce the number of package hallucinations while maintaining code quality. Our experiments and findings highlight package hallucinations as a persistent and systemic phenomenon while using state-of-the-art LLMs for code generation, and a significant challenge which deserves the research community's urgent attention.
LLM Ensemble -- which involves the comprehensive use of multiple large language models (LLMs), each aimed at handling user queries during downstream inference, to benefit from their individual strengths -- has gained substantial attention recently. The widespread availability of LLMs, coupled with their varying strengths and out-of-the-box usability, has profoundly advanced the field of LLM Ensemble. This paper presents the first systematic review of recent developments in LLM Ensemble. First, we introduce our taxonomy of LLM Ensemble and discuss several related research problems. Then, we provide a more in-depth classification of the methods under the broad categories of "ensemble-before-inference, ensemble-during-inference, ensemble-after-inference'', and review all relevant methods. Finally, we introduce related benchmarks and applications, summarize existing studies, and suggest several future research directions. A curated list of papers on LLM Ensemble is available at https://github.com/junchenzhi/Awesome-LLM-Ensemble.
Significant progress has been made in automated problem-solving using societies of agents powered by large language models (LLMs). In finance, efforts have largely focused on single-agent systems handling specific tasks or multi-agent frameworks independently gathering data. However, multi-agent systems' potential to replicate real-world trading firms' collaborative dynamics remains underexplored. TradingAgents proposes a novel stock trading framework inspired by trading firms, featuring LLM-powered agents in specialized roles such as fundamental analysts, sentiment analysts, technical analysts, and traders with varied risk profiles. The framework includes Bull and Bear researcher agents assessing market conditions, a risk management team monitoring exposure, and traders synthesizing insights from debates and historical data to make informed decisions. By simulating a dynamic, collaborative trading environment, this framework aims to improve trading performance. Detailed architecture and extensive experiments reveal its superiority over baseline models, with notable improvements in cumulative returns, Sharpe ratio, and maximum drawdown, highlighting the potential of multi-agent LLM frameworks in financial trading. TradingAgents is available at https://github.com/PioneerFintech.
Automatic LLM benchmarks, such as AlpacaEval 2.0, Arena-Hard-Auto, and MT-Bench, have become popular for evaluating language models due to their cost-effectiveness and scalability compared to human evaluation. Achieving high win rates on these benchmarks can significantly boost the promotional impact of newly released language models. This promotional benefit may motivate tricks, such as manipulating model output length or style to game win rates, even though several mechanisms have been developed to control length and disentangle style to reduce gameability. Nonetheless, we show that even a "null model" that always outputs a constant response (irrelevant to input instructions) can cheat automatic benchmarks and achieve top-ranked win rates: an 86.5% LC win rate on AlpacaEval 2.0; an 83.0 score on Arena-Hard-Auto; and a 9.55 score on MT-Bench. Moreover, the crafted cheating outputs are transferable because we assume that the instructions of these benchmarks (e.g., 805 samples of AlpacaEval 2.0) are private and cannot be accessed. While our experiments are primarily proof-of-concept, an adversary could use LLMs to generate more imperceptible cheating responses, unethically benefiting from high win rates and promotional impact. Our findings call for the development of anti-cheating mechanisms for reliable automatic benchmarks. The code is available at https://github.com/sail-sg/Cheating-LLM-Benchmarks.
Network device and system health management is the foundation of modern network operations and maintenance. Traditional health management methods, relying on expert identification or simple rule-based algorithms, struggle to cope with the dynamic heterogeneous networks (DHNs) environment. Moreover, current state-of-the-art distributed anomaly detection methods, which utilize specific machine learning techniques, lack multi-scale adaptivity for heterogeneous device information, resulting in unsatisfactory diagnostic accuracy for DHNs. In this paper, we develop an LLM-assisted end-to-end intelligent network health management framework. The framework first proposes a Multi-Scale Semanticized Anomaly Detection Model (MSADM), incorporating semantic rule trees with an attention mechanism to address the multi-scale anomaly detection problem in DHNs. Secondly, a chain-of-thought-based large language model is embedded in downstream to adaptively analyze the fault detection results and produce an analysis report with detailed fault information and optimization strategies. Experimental results show that the accuracy of our proposed MSADM for heterogeneous network entity anomaly detection is as high as 91.31\%.
Recently, utilizing large language models (LLMs) for metaphor detection has achieved promising results. However, these methods heavily rely on the capabilities of closed-source LLMs, which come with relatively high inference costs and latency. To address this, we propose a method for metaphor detection by fine-tuning open-source LLMs, effectively reducing inference costs and latency with a single inference step. Furthermore, metaphor detection suffers from a severe data scarcity problem, which hinders effective fine-tuning of LLMs. To tackle this, we introduce Curriculum-style Data Augmentation (CDA). Specifically, before fine-tuning, we evaluate the training data to identify correctly predicted instances for fine-tuning, while incorrectly predicted instances are used as seed data for data augmentation. This approach enables the model to quickly learn simpler knowledge and progressively acquire more complex knowledge, thereby improving performance incrementally. Experimental results demonstrate that our method achieves state-of-the-art performance across all baselines. Additionally, we provide detailed ablation studies to validate the effectiveness of CDA.
To enhance the reasoning capabilities of large language models (LLMs), self-consistency has gained significant popularity by combining multiple sampling with majority voting. However, the state-of-the-art self-consistency approaches consume substantial computational resources and lead to significant additional time costs due to the multiple sampling. This prevents its full potential from being realized in scenarios where computational resources are critical. To improve the inference efficiency, this paper introduces \textit{path-consistency}, a method that leverages the confidence of answers generated in earlier branches to identify the prefix of the most promising path. By dynamically guiding the generation of subsequent branches based on this prefix, the \textit{path-consistency} mitigates both the errors and redundancies from random or less useful sampling in self-consistency. As a result, it can significantly accelerate the inference process by reducing the number of tokens generated. Our extensive empirical evaluation shows that the \textit{path-consistency} achieves significant acceleration in inference latency ranging from $7.8\%$ to $40.5\%$, while maintaining or even improving task accuracy across different datasets, including mathematical reasoning, common sense reasoning, symbolic reasoning, and code generation.
The widespread applications of large language models (LLMs) have brought about concerns regarding their potential misuse. Although aligned with human preference data before release, LLMs remain vulnerable to various malicious attacks. In this paper, we adopt a red-teaming strategy to enhance LLM safety and introduce SeqAR, a simple yet effective framework to design jailbreak prompts automatically. The SeqAR framework generates and optimizes multiple jailbreak characters and then applies sequential jailbreak characters in a single query to bypass the guardrails of the target LLM. Different from previous work which relies on proprietary LLMs or seed jailbreak templates crafted by human expertise, SeqAR can generate and optimize the jailbreak prompt in a cold-start scenario using open-sourced LLMs without any seed jailbreak templates. Experimental results show that SeqAR achieves attack success rates of 88% and 60% in bypassing the safety alignment of GPT-3.5-1106 and GPT-4, respectively. Furthermore, we extensively evaluate the transferability of the generated templates across different LLMs and held-out malicious requests, while also exploring defense strategies against the jailbreak attack designed by SeqAR.
Efficient finetuning of large language models (LLMs) aims to adapt the LLMs with reduced computational and memory cost. Previous LoRA-based approaches initialize the low-rank matrices with Gaussian distribution and zero values while keeping the original weight matrices frozen. However, the trainable model parameters optimized in an unguided subspace might interfere with the well-learned subspace of the pretrained weight matrices. In this paper, we propose MiLoRA, a simple yet effective LLM finetuning approach that only updates the minor singular components of the weight matrix while keeping the principal singular components frozen. It is observed that the minor matrix corresponds to the noisy or long-tail information, while the principal matrix contains important knowledge. The MiLoRA initializes the low-rank matrices within a subspace that is orthogonal to the principal matrix, thus the pretrained knowledge is expected to be well preserved. During finetuning, MiLoRA makes the most use of the less-optimized subspace for learning the labeled dataset. Extensive experiments on commonsense reasoning, math reasoning, instruction following and visual instruction following benchmarks present the superior performance of our method.
Existing tools for laypeople to create personal classifiers often assume a motivated user working uninterrupted in a single, lengthy session. However, users tend to engage with social media casually, with many short sessions on an ongoing, daily basis. To make creating personal classifiers for content curation easier for such users, tools should support rapid initialization and iterative refinement. In this work, we compare three strategies -- (1) example labeling, (2) rule writing, and (3) large language model (LLM) prompting -- for end users to build personal content classifiers. From an experiment with 37 non-programmers tasked with creating personalized moderation filters, we found that participants preferred different initializing strategies in different contexts, despite LLM prompting's better performance. However, all strategies faced challenges with iterative refinement. To overcome challenges in iterating on their prompts, participants even adopted hybrid approaches such as providing examples as in-context examples or writing rule-like prompts.
Large Language Models (LLMs) are widely used for tasks such as natural language and code generation, but their outputs often suffer from issues like hallucination, toxicity, and incorrect results. Current libraries for structured LLM generation rely on left-to-right decoding without support for backtracking, limiting the ability to correct or refine outputs mid-generation. To address this, we introduce IterGen, a user-friendly library for iterative, grammar-guided LLM generation that enables users to move both forward and backward within the generated output based on grammar symbols. By leveraging a symbol-to-position mapping and maintaining the key-value (KV) cache state, IterGen ensures efficient and structured generation while allowing for corrections during the process. We demonstrate IterGen's effectiveness in two important applications: reducing privacy leakage in LLM outputs and improving the accuracy of LLM-generated SQL and Vega-Lite queries. Our code and additional resources are available at https://structuredllm.com.
Advancements in large language models (LLMs) are sparking a proliferation of LLM-powered user experiences (UX). In product teams, designers often craft UX to meet user needs, but it is unclear how they engage with LLMs as a novel design material. Through a formative study with 12 designers, we find that designers seek a translational process that enables design requirements to shape and be shaped by LLM behavior, motivating a need for designerly adaptation to facilitate this translation. We then built Canvil, a Figma widget that operationalizes designerly adaptation. We used Canvil as a probe to study designerly adaptation in a group-based design study (6 groups, N=17), finding that designers constructively iterated on both adaptation approaches and interface designs to enhance end-user interaction with LLMs. Furthermore, designers identified promising collaborative workflows for designerly adaptation. Our work opens new avenues for processes and tools that foreground designers' human-centered expertise when developing LLM-powered applications.
Large Language Models (LLMs) heavily rely on high-quality training data, making data valuation crucial for optimizing model performance, especially when working within a limited budget. In this work, we aim to offer a third-party data valuation approach that benefits both data providers and model developers. We introduce a linearized future influence kernel (LinFiK), which assesses the value of individual data samples in improving LLM performance during training. We further propose ALinFiK, a learning strategy to approximate LinFiK, enabling scalable data valuation. Our comprehensive evaluations demonstrate that this approach surpasses existing baselines in effectiveness and efficiency, demonstrating significant scalability advantages as LLM parameters increase.
Time series analysis provides essential insights for real-world system dynamics and informs downstream decision-making, yet most existing methods often overlook the rich contextual signals present in auxiliary modalities. To bridge this gap, we introduce TimeXL, a multi-modal prediction framework that integrates a prototype-based time series encoder with three collaborating Large Language Models (LLMs) to deliver more accurate predictions and interpretable explanations. First, a multi-modal prototype-based encoder processes both time series and textual inputs to generate preliminary forecasts alongside case-based rationales. These outputs then feed into a prediction LLM, which refines the forecasts by reasoning over the encoder's predictions and explanations. Next, a reflection LLM compares the predicted values against the ground truth, identifying textual inconsistencies or noise. Guided by this feedback, a refinement LLM iteratively enhances text quality and triggers encoder retraining. This closed-loop workflow -- prediction, critique (reflect), and refinement -- continuously boosts the framework's performance and interpretability. Empirical evaluations on four real-world datasets demonstrate that TimeXL achieves up to 8.9\% improvement in AUC and produces human-centric, multi-modal explanations, highlighting the power of LLM-driven reasoning for time series prediction.
Recent advances in large language models (LLMs) have demonstrated their potential as planners in human-robot collaboration (HRC) scenarios, offering a promising alternative to traditional planning methods. LLMs, which can generate structured plans by reasoning over natural language inputs, have the ability to generalize across diverse tasks and adapt to human instructions. This paper investigates the potential of LLMs to facilitate planning in the context of human-robot collaborative tasks, with a focus on their ability to reason from high-level, vague human inputs, and fine-tune plans based on real-time feedback. We propose a novel hybrid framework that combines LLMs with human feedback to create dynamic, context-aware task plans. Our work also highlights how a single, concise prompt can be used for a wide range of tasks and environments, overcoming the limitations of long, detailed structured prompts typically used in prior studies. By integrating user preferences into the planning loop, we ensure that the generated plans are not only effective but aligned with human intentions.
Large Language Models (LLMs) have demonstrated tremendous potential as the next-generation ranking-based recommendation system. Many recent works have shown that LLMs can significantly outperform conventional click-through-rate (CTR) prediction approaches. Despite such promising results, the computational inefficiency inherent in the current training paradigm makes it particularly challenging to train LLMs for ranking-based recommendation tasks on large datasets. To train LLMs for CTR prediction, most existing studies adopt the prevalent ''sliding-window'' paradigm. Given a sequence of $m$ user interactions, a unique training prompt is constructed for each interaction by designating it as the prediction target along with its preceding $n$ interactions serving as context. In turn, the sliding-window paradigm results in an overall complexity of $O(mn^2)$ that scales linearly with the length of user interactions. Consequently, a direct adoption to train LLMs with such strategy can result in prohibitively high training costs as the length of interactions grows. To alleviate the computational inefficiency, we propose a novel training paradigm, namely Dynamic Target Isolation (DTI), that structurally parallelizes the training of $k$ (where $k >> 1$) target interactions. Furthermore, we identify two major bottlenecks - hidden-state leakage and positional bias overfitting - that limit DTI to only scale up to a small value of $k$ (e.g., 5) then propose a computationally light solution to effectively tackle each. Through extensive experiments on three widely adopted public CTR datasets, we empirically show that DTI reduces training time by an average of $\textbf{92%}$ (e.g., from $70.5$ hrs to $5.31$ hrs), without compromising CTR prediction performance.
Structure reasoning is a fundamental capability of large language models (LLMs), enabling them to reason about structured commonsense and answer multi-hop questions. However, existing benchmarks for structure reasoning mainly focus on horizontal and coordinate structures (\emph{e.g.} graphs), overlooking the hierarchical relationships within them. Hierarchical structure reasoning is crucial for human cognition, particularly in memory organization and problem-solving. It also plays a key role in various real-world tasks, such as information extraction and decision-making. To address this gap, we propose HiBench, the first framework spanning from initial structure generation to final proficiency assessment, designed to benchmark the hierarchical reasoning capabilities of LLMs systematically. HiBench encompasses six representative scenarios, covering both fundamental and practical aspects, and consists of 30 tasks with varying hierarchical complexity, totaling 39,519 queries. To evaluate LLMs comprehensively, we develop five capability dimensions that depict different facets of hierarchical structure understanding. Through extensive evaluation of 20 LLMs from 10 model families, we reveal key insights into their capabilities and limitations: 1) existing LLMs show proficiency in basic hierarchical reasoning tasks; 2) they still struggle with more complex structures and implicit hierarchical representations, especially in structural modification and textual reasoning. Based on these findings, we create a small yet well-designed instruction dataset, which enhances LLMs' performance on HiBench by an average of 88.84\% (Llama-3.1-8B) and 31.38\% (Qwen2.5-7B) across all tasks. The HiBench dataset and toolkit are available here, https://github.com/jzzzzh/HiBench, to encourage evaluation.
Large Language Models (LLMs) have been observed to process non-human-readable text sequences, such as jailbreak prompts, often viewed as a bug for aligned LLMs. In this work, we present a systematic investigation challenging this perception, demonstrating that unnatural languages - strings that appear incomprehensible to humans but maintain semantic meanings for LLMs - contain latent features usable by models. Notably, unnatural languages possess latent features that can be generalized across different models and tasks during inference. Furthermore, models fine-tuned on unnatural versions of instruction datasets perform on-par with those trained on natural language, achieving 49.71 win rates in Length-controlled AlpacaEval 2.0 in average across various base models. In addition, through comprehensive analysis, we demonstrate that LLMs process unnatural languages by filtering noise and inferring contextual meaning from filtered words.
While large language models (LLMs) are increasingly used to assist users in various tasks through natural language interactions, these interactions often fall short due to LLMs' limited ability to infer contextual nuances and user intentions, unlike humans. To address this challenge, we draw inspiration from the Gricean Maxims--human communication theory that suggests principles of effective communication--and aim to derive design insights for enhancing human-AI interactions (HAI). Through participatory design workshops with communication experts, designers, and end-users, we identified ways to apply these maxims across the stages of the HAI cycle. Our findings include reinterpreted maxims tailored to human-LLM contexts and nine actionable design considerations categorized by interaction stage. These insights provide a concrete framework for designing more cooperative and user-centered LLM-based systems, bridging theoretical foundations in communication with practical applications in HAI.
Despite significant advancements in Large Language Models (LLMs), developing advanced reasoning capabilities in LLMs remains a key challenge. Process Reward Models (PRMs) have demonstrated exceptional promise in enhancing reasoning by providing step-wise feedback, particularly in the context of mathematical reasoning. However, their application to broader reasoning domains remains understudied, largely due to the high costs associated with manually creating step-level supervision. In this work, we explore the potential of PRMs in graph reasoning problems - a domain that demands sophisticated multi-step reasoning and offers opportunities for automated step-level data generation using established graph algorithms. We introduce GraphSILO, the largest dataset for graph reasoning problems with fine-grained step-wise labels, built using automated Task-oriented Trajectories and Monte Carlo Tree Search (MCTS) to generate detailed reasoning steps with step-wise labels. Building upon this dataset, we train GraphPRM, the first PRM designed for graph reasoning problems, and evaluate its effectiveness in two key settings: inference-time scaling and reinforcement learning via Direct Preference Optimization (DPO). Experimental results show that GraphPRM significantly improves LLM performance across 13 graph reasoning tasks, delivering a 9% gain for Qwen2.5-7B and demonstrating transferability to new graph reasoning datasets and new reasoning domains like mathematical problem-solving. Notably, GraphPRM enhances LLM performance on GSM8K and Math500, underscoring the cross-domain applicability of graph-based reasoning rewards. Our findings highlight the potential of PRMs in advancing reasoning across diverse domains, paving the way for more versatile and effective LLMs.
Federated learning systems have been identified as an efficient approach to scaling distributed model training with a large amount of participants or data owners while guaranteeing data privacy. To apply the current most popular pre-trained large language models to other domains with data privacy guarantee requirements, existing works propose fine-tuning the pre-trained large language models in federated learning environments across data owners using the parameter efficient fine-tuning approaches, LoRA. To address the resource and data heterogeneous issues for the participants, previous works adopted heterogeneous LoRA using different ranks for different clients and pending their rank, which brings bias for the parameter aggregation. To address this issue, we propose HLoRA, an efficient federated learning system utilizing a modified LoRA approach that incorporates rank heterogeneity to optimize communication and computational efficiency. Experimental results, conducted using the Microsoft Research Paraphrase Corpus (MRPC), Quora Question Pairs (QQP) and Recognizing Textual Entailment (RTE), within the Plato federated learning framework, demonstrate that our method not only reduces resource demands but also outperforms traditional LoRA applications in terms of convergence speed and final model accuracy. This study shows that our approach can significantly improve the practical deployment of federated LLM fine-tuning, particularly in environments with diverse client resources.
Large language models (LLMs) often struggle with knowledge intensive NLP tasks, such as answering "Who won the latest World Cup?" because the knowledge they learn during training may be insufficient or outdated. Conditioning generation on retrieved documents -- a technique known as retrieval augmented generation (RAG) -- mitigates these shortcomings by allowing the model to leverage in-context information. Practitioners can improve LLM RAG performance by fine-tuning on retrieval-augmented instructions, but must beware that this can cause undesirable model behaviors like hallucinations. We attribute this degradation to the fact that the training data is likely to be out-of-distribution for the model and may suffer from quality issues, such as misalignment between retrievals and target responses (since retrievals are frequently added post-hoc). We propose a recipe for training RAG-enabled LLMs using self-generated demonstrations, thereby avoiding training on out-of-distribution text and integrating retrievals into the LLM responses. We evaluate our method on knowledge intensive question answering (QA) tasks and show that our method teaches LLMs to properly handle in-context retrievals and abstain from questions it will likely get wrong. Compared to conventional RA-IT methods, our method prevents model degradation in non-RAG settings while exhibiting superior QA performance.
The development of large language models (LLMs) has raised concerns about potential misuse. One practical solution is to embed a watermark in the text, allowing ownership verification through watermark extraction. Existing methods primarily focus on defending against modification attacks, often neglecting other spoofing attacks. For example, attackers can alter the watermarked text to produce harmful content without compromising the presence of the watermark, which could lead to false attribution of this malicious content to the LLM. This situation poses a serious threat to the LLMs service providers and highlights the significance of achieving modification detection and generated-text detection simultaneously. Therefore, we propose a technique to detect modifications in text for unbiased watermark which is sensitive to modification. We introduce a new metric called ``discarded tokens", which measures the number of tokens not included in watermark detection. When a modification occurs, this metric changes and can serve as evidence of the modification. Additionally, we improve the watermark detection process and introduce a novel method for unbiased watermark. Our experiments demonstrate that we can achieve effective dual detection capabilities: modification detection and generated-text detection by watermark.
Credit assignment, the process of attributing credit or blame to individual agents for their contributions to a team's success or failure, remains a fundamental challenge in multi-agent reinforcement learning (MARL), particularly in environments with sparse rewards. Commonly-used approaches such as value decomposition often lead to suboptimal policies in these settings, and designing dense reward functions that align with human intuition can be complex and labor-intensive. In this work, we propose a novel framework where a large language model (LLM) generates dense, agent-specific rewards based on a natural language description of the task and the overall team goal. By learning a potential-based reward function over multiple queries, our method reduces the impact of ranking errors while allowing the LLM to evaluate each agent's contribution to the overall task. Through extensive experiments, we demonstrate that our approach achieves faster convergence and higher policy returns compared to state-of-the-art MARL baselines.
Objective: This study aims to develop and validate an evaluation framework to ensure the safety and reliability of mental health chatbots, which are increasingly popular due to their accessibility, human-like interactions, and context-aware support. Materials and Methods: We created an evaluation framework with 100 benchmark questions and ideal responses, and five guideline questions for chatbot responses. This framework, validated by mental health experts, was tested on a GPT-3.5-turbo-based chatbot. Automated evaluation methods explored included large language model (LLM)-based scoring, an agentic approach using real-time data, and embedding models to compare chatbot responses against ground truth standards. Results: The results highlight the importance of guidelines and ground truth for improving LLM evaluation accuracy. The agentic method, dynamically accessing reliable information, demonstrated the best alignment with human assessments. Adherence to a standardized, expert-validated framework significantly enhanced chatbot response safety and reliability. Discussion: Our findings emphasize the need for comprehensive, expert-tailored safety evaluation metrics for mental health chatbots. While LLMs have significant potential, careful implementation is necessary to mitigate risks. The superior performance of the agentic approach underscores the importance of real-time data access in enhancing chatbot reliability. Conclusion: The study validated an evaluation framework for mental health chatbots, proving its effectiveness in improving safety and reliability. Future work should extend evaluations to accuracy, bias, empathy, and privacy to ensure holistic assessment and responsible integration into healthcare. Standardized evaluations will build trust among users and professionals, facilitating broader adoption and improved mental health support through technology.
This paper proposes a novel backdoor threat attacking the LLM-as-a-Judge evaluation regime, where the adversary controls both the candidate and evaluator model. The backdoored evaluator victimizes benign users by unfairly assigning inflated scores to adversary. A trivial single token backdoor poisoning 1% of the evaluator training data triples the adversary's score with respect to their legitimate score. We systematically categorize levels of data access corresponding to three real-world settings, (1) web poisoning, (2) malicious annotator, and (3) weight poisoning. These regimes reflect a weak to strong escalation of data access that highly correlates with attack severity. Under the weakest assumptions - web poisoning (1), the adversary still induces a 20% score inflation. Likewise, in the (3) weight poisoning regime, the stronger assumptions enable the adversary to inflate their scores from 1.5/5 to 4.9/5. The backdoor threat generalizes across different evaluator architectures, trigger designs, evaluation tasks, and poisoning rates. By poisoning 10% of the evaluator training data, we control toxicity judges (Guardrails) to misclassify toxic prompts as non-toxic 89% of the time, and document reranker judges in RAG to rank the poisoned document first 97% of the time. LLM-as-a-Judge is uniquely positioned at the intersection of ethics and technology, where social implications of mislead model selection and evaluation constrain the available defensive tools. Amidst these challenges, model merging emerges as a principled tool to offset the backdoor, reducing ASR to near 0% whilst maintaining SOTA performance. Model merging's low computational cost and convenient integration into the current LLM Judge training pipeline position it as a promising avenue for backdoor mitigation in the LLM-as-a-Judge setting.
Hallucinations in LLMs pose a significant concern to their safe deployment in real-world applications. Recent approaches have leveraged the latent space of LLMs for hallucination detection, but their embeddings, optimized for linguistic coherence rather than factual accuracy, often fail to clearly separate truthful and hallucinated content. To this end, we propose the Truthfulness Separator Vector (TSV), a lightweight and flexible steering vector that reshapes the LLM's representation space during inference to enhance the separation between truthful and hallucinated outputs, without altering model parameters. Our two-stage framework first trains TSV on a small set of labeled exemplars to form compact and well-separated clusters. It then augments the exemplar set with unlabeled LLM generations, employing an optimal transport-based algorithm for pseudo-labeling combined with a confidence-based filtering process. Extensive experiments demonstrate that TSV achieves state-of-the-art performance with minimal labeled data, exhibiting strong generalization across datasets and providing a practical solution for real-world LLM applications.
Personalized interaction is highly valued by parents in their story-reading activities with children. While AI-empowered story-reading tools have been increasingly used, their abilities to support personalized interaction with children are still limited. Recent advances in large language models (LLMs) show promise in facilitating personalized interactions, but little is known about how to effectively and appropriately use LLMs to enhance children's personalized story-reading experiences. This work explores this question through a design-based study. Drawing on a formative study, we designed and developed StoryMate, an LLM-empowered personalized interactive story-reading tool for children, following an empirical study with children, parents, and education experts. Our participants valued the personalized features in StoryMate, and also highlighted the need to support personalized content, guiding mechanisms, reading context variations, and interactive interfaces. Based on these findings, we propose a series of design recommendations for better using LLMs to empower children's personalized story reading and interaction.
The adaptivity and maneuvering capabilities of Autonomous Underwater Vehicles (AUVs) have drawn significant attention in oceanic research, due to the unpredictable disturbances and strong coupling among the AUV's degrees of freedom. In this paper, we developed large language model (LLM)-enhanced reinforcement learning (RL)-based adaptive S-surface controller for AUVs. Specifically, LLMs are introduced for the joint optimization of controller parameters and reward functions in RL training. Using multi-modal and structured explicit task feedback, LLMs enable joint adjustments, balance multiple objectives, and enhance task-oriented performance and adaptability. In the proposed controller, the RL policy focuses on upper-level tasks, outputting task-oriented high-level commands that the S-surface controller then converts into control signals, ensuring cancellation of nonlinear effects and unpredictable external disturbances in extreme sea conditions. Under extreme sea conditions involving complex terrain, waves, and currents, the proposed controller demonstrates superior performance and adaptability in high-level tasks such as underwater target tracking and data collection, outperforming traditional PID and SMC controllers.
Autonomous Vehicles (AVs) have entered the commercialization stage, but their limited ability to interact and express intentions still poses challenges in interactions with Human-driven Vehicles (HVs). Recent advances in large language models (LLMs) enable bidirectional human-machine communication, but the conflict between slow inference speed and the need for real-time decision-making challenges practical deployment. To address these issues, this paper introduces a parallel Actor-Reasoner framework designed to enable explicit bidirectional AV-HV interactions across multiple scenarios. First, by facilitating interactions between the LLM-driven Reasoner and heterogeneous simulated HVs during training, an interaction memory database, referred to as the Actor, is established. Then, by introducing the memory partition module and the two-layer memory retrieval module, the Actor's ability to handle heterogeneous HVs is significantly enhanced. Ablation studies and comparisons with other decision-making methods demonstrate that the proposed Actor-Reasoner framework significantly improves safety and efficiency. Finally, with the combination of the external Human-Machine Interface (eHMI) information derived from Reasoner's reasoning and the feasible action solutions retrieved from the Actor, the effectiveness of the proposed Actor-Reasoner is confirmed in multi-scenario field interactions. Our code is available at https://github.com/FanGShiYuu/Actor-Reasoner.
This tutorial presents a comprehensive introduction to Speculative Decoding (SD), an advanced technique for LLM inference acceleration that has garnered significant research interest in recent years. SD is introduced as an innovative decoding paradigm to mitigate the high inference latency stemming from autoregressive decoding in LLMs. At each decoding step, SD efficiently drafts several future tokens and then verifies them in parallel. This approach, unlike traditional autoregressive decoding, facilitates the simultaneous decoding of multiple tokens per step, thereby achieving promising 2x-4x speedups in LLM inference while maintaining original distributions. This tutorial delves into the latest techniques in SD, including draft model architectures and verification strategies. Additionally, it explores the acceleration potential and future research directions in this promising field. We aim for this tutorial to elucidate the current research landscape and offer insights for researchers interested in Speculative Decoding, ultimately contributing to more efficient LLM inference.
Processing long contexts has become a critical capability for modern large language models (LLMs). However, serving long-context LLMs comes with significant inference costs due to the high memory overhead of the key-value (KV) cache. Existing work leverages dynamic sparse attention algorithms (DSAes) to mitigate the KV cache overhead, but these algorithms rely on top-$k$ KV cache selection, which results in a trade-off between accuracy and efficiency. A larger $k$ improves accuracy but decreases efficiency, while a smaller $k$ boosts efficiency but compromises accuracy. To overcome this trade-off, this paper presents PSA, a $\underline{P}$rogressive $\underline{S}$parse $\underline{A}$ttention mechanism that integrates algorithmic innovations with system co-design to achieve both high inference accuracy and improved efficiency in LLM serving. The PSA algorithm adaptively adjusts the KV cache budget of different tokens and layers according to their real attention weight distributions, rather than relying on a fixed budget $k$. This enables high accuracy while minimizing KV cache usage. To further enhance execution efficiency, we introduce a pipelined iteration scheme that reduces CPU-GPU interleaving and synchronization overhead during PSA computation. Additionally, we implement unified GPU memory management that optimizes PSA's memory utilization by accounting for uneven memory requirements across different model layers. Extensive experimental results demonstrate that PSA reduces KV cache usage for attention computation by up to 2.4$\times$ and 8.8$\times$, and increases end-to-end serving throughput by up to 1.4$\times$ and 2.0$\times$, compared to state-of-the-art DSAes and systems without sparse attention, respectively.
Recent advancements in Large Language Models (LLMs) have expanded their context windows to unprecedented lengths, sparking debates about the necessity of Retrieval-Augmented Generation (RAG). To address the fragmented evaluation paradigms and limited cases in existing Needle-in-a-Haystack (NIAH), this paper introduces U-NIAH, a unified framework that systematically compares LLMs and RAG methods in controlled long context settings. Our framework extends beyond traditional NIAH by incorporating multi-needle, long-needle, and needle-in-needle configurations, along with different retrieval settings, while leveraging the synthetic Starlight Academy dataset-a fictional magical universe-to eliminate biases from pre-trained knowledge. Through extensive experiments, we investigate three research questions: (1) performance trade-offs between LLMs and RAG, (2) error patterns in RAG, and (3) RAG's limitations in complex settings. Our findings show that RAG significantly enhances smaller LLMs by mitigating the "lost-in-the-middle" effect and improving robustness, achieving an 82.58% win-rate over LLMs. However, we observe that retrieval noise and reverse chunk ordering degrade performance, while surprisingly, advanced reasoning LLMs exhibit reduced RAG compatibility due to sensitivity to semantic distractors. We identify typical error patterns including omission due to noise, hallucination under high noise critical condition, and self-doubt behaviors. Our work not only highlights the complementary roles of RAG and LLMs, but also provides actionable insights for optimizing deployments. Code: https://github.com/Tongji-KGLLM/U-NIAH.
How human cognitive abilities are formed has long captivated researchers. However, a significant challenge lies in developing meaningful methods to measure these complex processes. With the advent of large language models (LLMs), which now rival human capabilities in various domains, we are presented with a unique testbed to investigate human cognition through a new lens. Among the many facets of cognition, one particularly crucial aspect is the concept of semantic size, the perceived magnitude of both abstract and concrete words or concepts. This study seeks to investigate whether LLMs exhibit similar tendencies in understanding semantic size, thereby providing insights into the underlying mechanisms of human cognition. We begin by exploring metaphorical reasoning, comparing how LLMs and humans associate abstract words with concrete objects of varying sizes. Next, we examine LLMs' internal representations to evaluate their alignment with human cognitive processes. Our findings reveal that multi-modal training is crucial for LLMs to achieve more human-like understanding, suggesting that real-world, multi-modal experiences are similarly vital for human cognitive development. Lastly, we examine whether LLMs are influenced by attention-grabbing headlines with larger semantic sizes in a real-world web shopping scenario. The results show that multi-modal LLMs are more emotionally engaged in decision-making, but this also introduces potential biases, such as the risk of manipulation through clickbait headlines. Ultimately, this study offers a novel perspective on how LLMs interpret and internalize language, from the smallest concrete objects to the most profound abstract concepts like love. The insights gained not only improve our understanding of LLMs but also provide new avenues for exploring the cognitive abilities that define human intelligence.
The advent of Large Language Models (LLMs) has revolutionized natural language processing. However, these models face challenges in retrieving precise information from vast datasets. Retrieval-Augmented Generation (RAG) was developed to combining LLMs with external information retrieval systems to enhance the accuracy and context of responses. Despite improvements, RAG still struggles with comprehensive retrieval in high-volume, low-information-density databases and lacks relational awareness, leading to fragmented answers. To address this, this paper introduces the Pseudo-Knowledge Graph (PKG) framework, designed to overcome these limitations by integrating Meta-path Retrieval, In-graph Text and Vector Retrieval into LLMs. By preserving natural language text and leveraging various retrieval techniques, the PKG offers a richer knowledge representation and improves accuracy in information retrieval. Extensive evaluations using Open Compass and MultiHop-RAG benchmarks demonstrate the framework's effectiveness in managing large volumes of data and complex relationships.
Generalizing language-conditioned robotic policies to new tasks remains a significant challenge, hampered by the lack of suitable simulation benchmarks. In this paper, we address this gap by introducing GemBench, a novel benchmark to assess generalization capabilities of vision-language robotic manipulation policies. GemBench incorporates seven general action primitives and four levels of generalization, spanning novel placements, rigid and articulated objects, and complex long-horizon tasks. We evaluate state-of-the-art approaches on GemBench and also introduce a new method. Our approach 3D-LOTUS leverages rich 3D information for action prediction conditioned on language. While 3D-LOTUS excels in both efficiency and performance on seen tasks, it struggles with novel tasks. To address this, we present 3D-LOTUS++, a framework that integrates 3D-LOTUS's motion planning capabilities with the task planning capabilities of LLMs and the object grounding accuracy of VLMs. 3D-LOTUS++ achieves state-of-the-art performance on novel tasks of GemBench, setting a new standard for generalization in robotic manipulation. The benchmark, codes and trained models are available at https://www.di.ens.fr/willow/research/gembench/.
Large Language Models (LLMs) have recently shown remarkable advancement in various NLP tasks. As such, a popular trend has emerged lately where NLP researchers extract word/sentence/document embeddings from these large decoder-only models and use them for various inference tasks with promising results. However, it is still unclear whether the performance improvement of LLM-induced embeddings is merely because of scale or whether underlying embeddings they produce significantly differ from classical encoding models like Word2Vec, GloVe, Sentence-BERT (SBERT) or Universal Sentence Encoder (USE). This is the central question we investigate in the paper by systematically comparing classical decontextualized and contextualized word embeddings with the same for LLM-induced embeddings. Our results show that LLMs cluster semantically related words more tightly and perform better on analogy tasks in decontextualized settings. However, in contextualized settings, classical models like SimCSE often outperform LLMs in sentence-level similarity assessment tasks, highlighting their continued relevance for fine-grained semantics.
Large Language Models (LLMs) have recently shown remarkable advancement in various NLP tasks. As such, a popular trend has emerged lately where NLP researchers extract word/sentence/document embeddings from these large decoder-only models and use them for various inference tasks with promising results. However, it is still unclear whether the performance improvement of LLM-induced embeddings is merely because of scale or whether underlying embeddings they produce significantly differ from classical encoding models like Word2Vec, GloVe, Sentence-BERT (SBERT) or Universal Sentence Encoder (USE). This is the central question we investigate in the paper by systematically comparing classical decontextualized and contextualized word embeddings with the same for LLM-induced embeddings. Our results show that LLMs cluster semantically related words more tightly and perform better on analogy tasks in decontextualized settings. However, in contextualized settings, classical models like SimCSE often outperform LLMs in sentence-level similarity assessment tasks, highlighting their continued relevance for fine-grained semantics.
The development of Large Language Models (LLMs) often confronts challenges stemming from the heavy reliance on human annotators in the reinforcement learning with human feedback (RLHF) framework, or the frequent and costly external queries tied to the self-instruct paradigm. In this work, we pivot to Reinforcement Learning (RL) -- but with a twist. Diverging from the typical RLHF, which refines LLMs following instruction data training, we use RL to directly generate the foundational instruction dataset that alone suffices for fine-tuning. Our method, TeaMs-RL, uses a suite of textual operations and rules, prioritizing the diversification of training datasets. It facilitates the generation of high-quality data without excessive reliance on external advanced models, paving the way for a single fine-tuning step and negating the need for subsequent RLHF stages. Our findings highlight key advantages of our approach: reduced need for human involvement and fewer model queries (only 5.73% of the strong baseline's total), along with enhanced capabilities of LLMs in crafting and comprehending complex instructions compared to strong baselines, and substantially improved model privacy protection. Code is available at the link: https://github.com/SafeRL-Lab/TeaMs-RL
Large Language Models (LLMs) are susceptible to security and safety threats, such as prompt injection, prompt extraction, and harmful requests. One major cause of these vulnerabilities is the lack of an instruction hierarchy. Modern LLM architectures treat all inputs equally, failing to distinguish between and prioritize various types of instructions, such as system messages, user prompts, and data. As a result, lower-priority user prompts may override more critical system instructions, including safety protocols. Existing approaches to achieving instruction hierarchy, such as delimiters and instruction-based training, do not address this issue at the architectural level. We introduce the Instructional Segment Embedding (ISE) technique, inspired by BERT, to modern large language models, which embeds instruction priority information directly into the model. This approach enables models to explicitly differentiate and prioritize various instruction types, significantly improving safety against malicious prompts that attempt to override priority rules. Our experiments on the Structured Query and Instruction Hierarchy benchmarks demonstrate an average robust accuracy increase of up to 15.75% and 18.68%, respectively. Furthermore, we observe an improvement in instruction-following capability of up to 4.1% evaluated on AlpacaEval. Overall, our approach offers a promising direction for enhancing the safety and effectiveness of LLM architectures.
Low-resource languages face significant challenges due to the lack of sufficient linguistic data, resources, and tools for tasks such as supervised learning, annotation, and classification. This shortage hinders the development of accurate models and datasets, making it difficult to perform critical NLP tasks like sentiment analysis or hate speech detection. To bridge this gap, Large Language Models (LLMs) present an opportunity for potential annotators, capable of generating datasets and resources for these underrepresented languages. In this paper, we focus on Marathi, a low-resource language, and evaluate the performance of both closed-source and open-source LLMs as annotators, while also comparing these results with fine-tuned BERT models. We assess models such as GPT-4o and Gemini 1.0 Pro, Gemma 2 (2B and 9B), and Llama 3.1 (8B and 405B) on classification tasks including sentiment analysis, news classification, and hate speech detection. Our findings reveal that while LLMs excel in annotation tasks for high-resource languages like English, they still fall short when applied to Marathi. Even advanced models like GPT-4o and Llama 3.1 405B underperform compared to fine-tuned BERT-based baselines, with GPT-4o and Llama 3.1 405B trailing fine-tuned BERT by accuracy margins of 10.2% and 14.1%, respectively. This highlights the limitations of LLMs as annotators for low-resource languages.
The efficacy of detectors for texts generated by large language models (LLMs) substantially depends on the availability of large-scale training data. However, white-box zero-shot detectors, which require no such data, are limited by the accessibility of the source model of the LLM-generated text. In this paper, we propose a simple yet effective black-box zero-shot detection approach based on the observation that, from the perspective of LLMs, human-written texts typically contain more grammatical errors than LLM-generated texts. This approach involves calculating the Grammar Error Correction Score (GECScore) for the given text to differentiate between human-written and LLM-generated text. Experimental results show that our method outperforms current state-of-the-art (SOTA) zero-shot and supervised methods, achieving an average AUROC of 98.62% across XSum and Writing Prompts dataset. Additionally, our approach demonstrates strong reliability in the wild, exhibiting robust generalization and resistance to paraphrasing attacks. Data and code are available at: https://github.com/NLP2CT/GECScore.
The processing of numerical values is a rapidly developing area in the field of Language Models (LLMs). Despite numerous advancements achieved by previous research, significant challenges persist, particularly within the healthcare domain. This paper investigates the limitations of Transformer models in understanding numerical values. \textit{Objective:} this research aims to categorize numerical values extracted from medical documents into eight specific physiological categories using CamemBERT-bio. \textit{Methods:} In a context where scalable methods and Large Language Models (LLMs) are emphasized, we explore lifting the limitations of transformer-based models. We examine two strategies: fine-tuning CamemBERT-bio on a small medical dataset, integrating Label Embedding for Self-Attention (LESA), and combining LESA with additional enhancement techniques such as Xval. Given that CamemBERT-bio is already pre-trained on a large medical dataset, the first approach aims to update its encoder with the newly added label embeddings technique. In contrast, the second approach seeks to develop multiple representations of numbers (contextual and magnitude-based) to achieve more robust number embeddings. \textit{Results:} As anticipated, fine-tuning the standard CamemBERT-bio on our small medical dataset did not improve F1 scores. However, significant improvements were observed with CamemBERT-bio + LESA, resulting in an over 13\% increase. Similar enhancements were noted when combining LESA with Xval, outperforming conventional methods and giving comparable results to GPT-4 \textit{Conclusions and Novelty:} This study introduces two innovative techniques for handling numerical data, which are also applicable to other modalities. We illustrate how these techniques can improve the performance of Transformer-based models, achieving more reliable classification results even with small datasets.