Text_Generation
-
FraQAT: Quantization Aware Training with Fractional bits
2025-10-16 16:01:08Luca Morreale, Alberto Gil C. P. Ramos, Malcolm Chadwick, Mehid Noroozi, Ruchika Chavhan, Abhinav Mehrotra, Sourav BhattacharyaAbstract
State-of-the-art (SOTA) generative models have demonstrated impressive capabilities in image synthesis or text generation, often with a large capacity model. However, these large models cannot be deployed on smartphones due to the limited availability of on-board memory and computations. Quantization methods lower the precision of the model parameters, allowing for efficient computations, \eg, in \INT{8}. Although aggressive quantization addresses efficiency and memory constraints, preserving the quality of the model remains a challenge. To retain quality in previous aggressive quantization, we propose a new fractional bits quantization (\short) approach. The novelty is a simple yet effective idea: we progressively reduce the model's precision from 32 to 4 bits per parameter, and exploit the fractional bits during optimization to maintain high generation quality. We show that the \short{} yields improved quality on a variety of diffusion models, including SD3.5-Medium, Sana, \pixart, and FLUX.1-schnell, while achieving $4-7\%$ lower FiD than standard QAT. Finally, we deploy and run Sana on a Samsung S25U, which runs on the Qualcomm SM8750-AB Snapdragon 8 Elite Hexagon Tensor Processor (HTP).
Abstract (translated)
最先进的（SOTA）生成模型在图像合成或文本生成方面表现出令人印象深刻的能力，通常使用的是大规模模型。然而，由于手机上的内存和计算资源有限，这些大型模型无法部署到智能手机上。量化方法通过降低模型参数的精度来实现高效的计算，例如采用INT8格式。尽管激进的量化方法可以解决效率和内存限制问题，但保持模型质量仍是一个挑战。为了在之前的激进量化中保留质量，我们提出了一种新的分数位量化（简称FracBits）方法。其新颖之处在于一个简单而有效的想法：我们逐步将模型精度从32位减少到每参数4位，并利用分数位进行优化以保持高质量的生成效果。我们展示了FracBits在多种扩散模型上的改进，包括SD3.5-Medium、Sana、PixArt和FLUX.1-schnell，在FiD指标上比标准QAT低4-7%。最后，我们在三星S25U设备（搭载高通SM8750-AB骁龙8 Elite Hexagon Tensor Processor (HTP)）上部署并运行了Sana模型。 这一段翻译概述了一种新的分数位量化方法，旨在解决大型生成模型在手机等资源受限的设备上的部署问题。该方法通过逐步减少模型参数的精度，并利用分数位进行优化，在保持高质量的同时实现了更高效的计算。
URL
https://arxiv.org/abs/2510.14823
PDF
-
Qwen3Guard Technical Report
2025-10-16 04:00:18Haiquan Zhao, Chenhan Yuan, Fei Huang, Xiaomeng Hu, Yichang Zhang, An Yang, Bowen Yu, Dayiheng Liu, Jingren Zhou, Junyang Lin, Baosong Yang, Chen Cheng, Jialong Tang, Jiandong Jiang, Jianwei Zhang, Jijie Xu, Ming Yan, Minmin Sun, Pei Zhang, Pengjun Xie, Qiaoyu Tang, Qin Zhu, Rong Zhang, Shibin Wu, Shuo Zhang, Tao He, Tianyi Tang, Tingyu Xia, Wei Liao, Weizhou Shen, Wenbiao Yin, Wenmeng Zhou, Wenyuan Yu, Xiaobin Wang, Xiaodong Deng, Xiaodong Xu, Xinyu Zhang, Yang Liu, Yeqiu Li, Yi Zhang, Yong Jiang, Yu Wan, Yuxin ZhouAbstract
As large language models (LLMs) become more capable and widely used, ensuring the safety of their outputs is increasingly critical. Existing guardrail models, though useful in static evaluation settings, face two major limitations in real-world applications: (1) they typically output only binary "safe/unsafe" labels, which can be interpreted inconsistently across diverse safety policies, rendering them incapable of accommodating varying safety tolerances across domains; and (2) they require complete model outputs before performing safety checks, making them fundamentally incompatible with streaming LLM inference, thereby preventing timely intervention during generation and increasing exposure to harmful partial outputs. To address these challenges, we present Qwen3Guard, a series of multilingual safety guardrail models with two specialized variants: Generative Qwen3Guard, which casts safety classification as an instruction-following task to enable fine-grained tri-class judgments (safe, controversial, unsafe); and Stream Qwen3Guard, which introduces a token-level classification head for real-time safety monitoring during incremental text generation. Both variants are available in three sizes (0.6B, 4B, and 8B parameters) and support up to 119 languages and dialects, providing comprehensive, scalable, and low-latency safety moderation for global LLM deployments. Evaluated across English, Chinese, and multilingual benchmarks, Qwen3Guard achieves state-of-the-art performance in both prompt and response safety classification. All models are released under the Apache 2.0 license for public use.
Abstract (translated)
随着大型语言模型（LLM）的能力和使用范围不断扩大，确保其输出的安全性变得越来越重要。现有的防护模型虽然在静态评估环境中很有用，但在实际应用中面临两个主要限制：(1) 它们通常仅提供二元的“安全/不安全”标签，在面对多样的安全政策时可能导致解释上的不一致，无法适应不同领域中的不同安全容忍度；(2) 它们需要在完成模型输出之后才能进行安全性检查，这使得它们与流式LLM推理本质上是不兼容的，因此无法及时干预生成过程，并增加了暴露于有害部分输出的风险。为了解决这些问题，我们提出了Qwen3Guard，一系列多语言安全防护模型，包含两种专门化的变体：Generative Qwen3Guard将安全性分类作为一种指令跟随任务来实现细粒度三元判断（安全、有争议、不安全）；Stream Qwen3Guard则引入了一个令牌级别的分类头部以实现在增量文本生成过程中的实时安全监测。这两种变体分别提供0.6B、4B和8B参数大小的模型，并支持多达119种语言和方言，为全球LLM部署提供了全面、可扩展且低延迟的安全管理方案。Qwen3Guard在英语、中文以及多语言基准测试中，在提示和响应安全性分类方面均达到了最先进的性能水平。所有模型均以Apache 2.0许可证发布供公众使用。
URL
https://arxiv.org/abs/2510.14276
PDF
-
Generating Fair Consensus Statements with Social Choice on Token-Level MDPs
2025-10-15 21:23:18Carter Blair, Kate LarsonAbstract
Current frameworks for consensus statement generation with large language models lack the inherent structure needed to provide provable fairness guarantees when aggregating diverse free-form opinions. We model the task as a multi-objective, token-level Markov Decision Process (MDP), where each objective corresponds to an agent's preference. Token-level rewards for each agent are derived from their policy (e.g., a personalized language model). This approach utilizes the finding that such policies implicitly define optimal Q-functions, providing a principled way to quantify rewards at each generation step without a value function (Rafailov et al., 2024). This MDP formulation creates a formal structure amenable to analysis using principles from social choice theory. We propose two approaches grounded in social choice theory. First, we propose a stochastic generation policy guaranteed to be in the ex-ante core, extending core stability concepts from voting theory to text generation. This policy is derived from an underlying distribution over complete statements that maximizes proportional fairness (Nash Welfare). Second, for generating a single statement, we target the maximization of egalitarian welfare using search algorithms within the MDP framework. Empirically, experiments using language models to instantiate agent policies show that search guided by the egalitarian objective generates consensus statements with improved worst-case agent alignment compared to baseline methods, including the Habermas Machine (Tessler et al., 2024).
Abstract (translated)
当前用于大型语言模型共识声明生成的框架缺乏内在结构，无法在聚合不同自由形式的意见时提供可证明的公平性保证。我们把这个任务建模为一个多目标、基于标记级别的马尔科夫决策过程（MDP），每个目标对应于一个代理人的偏好。每一个代理人的令牌级别奖励是从他们的策略中推导出来的（例如，个性化语言模型）。这种方法利用了这样一个发现：这样的策略隐含地定义了最优的Q函数，提供了一种量化每一步生成过程中奖励的原则方法，而无需价值函数（Rafailov等人，2024年）。这种MDP结构使得应用社会选择理论原理进行分析成为可能。 我们提出了两种基于社会选择理论的方法。首先，我们提出了一种保证在事前核心内的随机生成策略，将投票理论中的核心稳定性概念扩展到文本生成中。该策略来源于最大化比例公平性的完整声明的基础分布（纳什福利）。其次，在生成单一声明时，我们在MDP框架内使用搜索算法来追求平等福利的最大化。 实验结果表明，当用语言模型实现代理人策略时，基于平等目标引导的搜索可以产生比基线方法更优的最差情况代理对齐共识声明，包括哈贝马斯机器（Tessler等人，2024年）。
URL
https://arxiv.org/abs/2510.14106
PDF
-
FreshTab: Sourcing Fresh Data for Table-to-Text Generation Evaluation
2025-10-15 14:31:44Krist\'yna Onderkov\'a, Ond\v{r}ej Pl\'atek, Zden\v{e}k Kasner, Ond\v{r}ej Du\v{s}ekAbstract
Table-to-text generation (insight generation from tables) is a challenging task that requires precision in analyzing the data. In addition, the evaluation of existing benchmarks is affected by contamination of Large Language Model (LLM) training data as well as domain imbalance. We introduce FreshTab, an on-the-fly table-to-text benchmark generation from Wikipedia, to combat the LLM data contamination problem and enable domain-sensitive evaluation. While non-English table-to-text datasets are limited, FreshTab collects datasets in different languages on demand (we experiment with German, Russian and French in addition to English). We find that insights generated by LLMs from recent tables collected by our method appear clearly worse by automatic metrics, but this does not translate into LLM and human evaluations. Domain effects are visible in all evaluations, showing that a~domain-balanced benchmark is more challenging.
Abstract (translated)
表格到文本生成（从表格中生成洞察）是一项需要精确数据分析的任务。此外，现有基准的评估受到大型语言模型（LLM）训练数据污染以及领域不平衡的影响。我们引入了FreshTab，这是一个基于维基百科实时生成表格到文本基准的数据集，以解决LLM数据污染问题，并实现领域敏感性评估。尽管非英语的表格到文本数据集有限，但FreshTab可以根据需求收集不同语言的数据集（我们在实验中使用了德语、俄语和法语，以及英语）。我们发现，从最近通过我们的方法收集的表格中生成的LLM洞察在自动评分指标上明显较差，但这并未反映在LLM和人类评估中。所有评估中的领域效应表明，一个领域的平衡基准更具挑战性。
URL
https://arxiv.org/abs/2510.13598
PDF
-
MADREC: A Multi-Aspect Driven LLM Agent for Explainable and Adaptive Recommendation
2025-10-15 10:03:29Jiin Park, Misuk KimAbstract
Recent attempts to integrate large language models (LLMs) into recommender systems have gained momentum, but most remain limited to simple text generation or static prompt-based inference, failing to capture the complexity of user preferences and real-world interactions. This study proposes the Multi-Aspect Driven LLM Agent MADRec, an autonomous LLM-based recommender that constructs user and item profiles by unsupervised extraction of multi-aspect information from reviews and performs direct recommendation, sequential recommendation, and explanation generation. MADRec generates structured profiles via aspect-category-based summarization and applies Re-Ranking to construct high-density inputs. When the ground-truth item is missing from the output, the Self-Feedback mechanism dynamically adjusts the inference criteria. Experiments across multiple domains show that MADRec outperforms traditional and LLM-based baselines in both precision and explainability, with human evaluation further confirming the persuasiveness of the generated explanations.
Abstract (translated)
近期将大型语言模型（LLMs）集成到推荐系统中的尝试逐渐增多，但大多数仍局限于简单的文本生成或静态提示推理，未能捕捉用户偏好和现实世界互动的复杂性。本研究提出了一个多方面驱动的LLM代理——MADRec（Multi-Aspect Driven LLM Agent），这是一种自主的基于LLM的推荐器。它通过无监督地从评论中提取多方面的信息来构建用户和物品档案，并执行直接推荐、顺序推荐和解释生成。 MADRec利用以方面类别为基础的摘要生成结构化档案，并应用重排序技术构造高密度输入。当真实目标项目缺失在输出结果时，自反馈机制会动态调整推理标准。跨多个领域的实验表明，MADRec在精度和可解释性方面均优于传统方法以及基于LLM的方法基准线，且人类评估进一步确认了生成解释的说服力。
URL
https://arxiv.org/abs/2510.13371
PDF
-
In-Distribution Steering: Balancing Control and Coherence in Language Model Generation
2025-10-15 08:31:37Arthur Vogels, Benjamin Wong, Yann Choho, Annabelle Blangero, Milan BhanAbstract
Activation steering methods control large language model (LLM) behavior by modifying internal activations at inference time. However, most existing activation steering methods rely on a fixed steering strength, leading to either insufficient control or unadapted intervention that degrades text plausibility and coherence. We introduce In-Distribution Steering (IDS), a novel method that adapts steering strength based on the input data distribution in representation space. IDS dynamically adjusts interventions according to how far a given input lies within the distribution, enabling adaptive intervention and generation stability during text generation. Experiments demonstrate that IDS achieves strong accuracy on classification tasks while producing coherent text without collapse, making IDS particularly well suited for real-world applications.
Abstract (translated)
激活导向方法通过在推理时修改大型语言模型（LLM）的内部激活来控制其行为。然而，大多数现有的激活导向方法依赖于固定的导向强度，这可能导致对模型行为的控制不足或过度干预，从而降低生成文本的真实性和连贯性。我们引入了一种新颖的方法——分布内导向（IDS），它根据输入数据在表示空间中的分布动态调整导向强度。IDS能够根据给定输入与训练数据分布的距离来调整干预措施，这使得在文本生成过程中可以实现自适应的干预和生成稳定性。 实验结果表明，IDS在分类任务中取得了优异的表现，并且在不牺牲连贯性的前提下保持了文本的稳定生成。因此，IDS特别适合于现实世界的应用场景。
URL
https://arxiv.org/abs/2510.13285
PDF
-
Beating Harmful Stereotypes Through Facts: RAG-based Counter-speech Generation
2025-10-14 09:20:01Greta Damo, Elena Cabrio, Serena VillataAbstract
Counter-speech generation is at the core of many expert activities, such as fact-checking and hate speech, to counter harmful content. Yet, existing work treats counter-speech generation as pure text generation task, mainly based on Large Language Models or NGO experts. These approaches show severe drawbacks due to the limited reliability and coherence in the generated countering text, and in scalability, respectively. To close this gap, we introduce a novel framework to model counter-speech generation as knowledge-wise text generation process. Our framework integrates advanced Retrieval-Augmented Generation (RAG) pipelines to ensure the generation of trustworthy counter-speech for 8 main target groups identified in the hate speech literature, including women, people of colour, persons with disabilities, migrants, Muslims, Jews, LGBT persons, and other. We built a knowledge base over the United Nations Digital Library, EUR-Lex and the EU Agency for Fundamental Rights, comprising a total of 32,792 texts. We use the MultiTarget-CONAN dataset to empirically assess the quality of the generated counter-speech, both through standard metrics (i.e., JudgeLM) and a human evaluation. Results show that our framework outperforms standard LLM baselines and competitive approach, on both assessments. The resulting framework and the knowledge base pave the way for studying trustworthy and sound counter-speech generation, in hate speech and beyond.
Abstract (translated)
反言论生成是许多专家活动的核心，例如事实核查和仇恨言论处理，旨在对抗有害内容。然而，现有的研究将反言论生成视为纯粹的文本生成任务，并主要依赖于大型语言模型或非政府组织（NGO）专家。这些方法因生成的反驳文本可靠性和连贯性有限以及在可扩展性方面的问题而表现出严重缺陷。 为弥补这一不足，我们引入了一种新颖框架，将反言论生成建模为基于知识的文本生成过程。我们的框架整合了先进的检索增强生成（RAG）管道，以确保针对仇恨言论文献中识别出的8个主要目标群体产生可信的反驳言论，这些群体包括女性、有色人种、残疾人、移民、穆斯林、犹太人、LGBT人士以及其他群体。我们在联合国数字图书馆、EUR-Lex和欧盟基本权利局的基础上构建了一个知识库，总计包含32,792条文本。 我们使用MultiTarget-CONAN数据集通过标准指标（即JudgeLM）和人工评估来实证检验生成的反言论质量。结果显示，我们的框架在两项评估中均优于传统的大型语言模型基线和其他竞争方法。 由此产生的框架和知识库为研究可信且合理的反言论生成铺平了道路，不仅限于仇恨言论领域。
URL
https://arxiv.org/abs/2510.12316
PDF
-
GOAT: A Training Framework for Goal-Oriented Agent with Tools
2025-10-14 07:14:50Hyunji Min, Sangwon Jung, Junyoung Sung, Dosung Lee, Leekyeung Han, Paul Hongsuck SeoAbstract
Large language models (LLMs) have recently been extended beyond traditional text generation to serve as interactive agents capable of using external tools based on user intent. However, current LLM agents still show limited ability to handle goal-oriented queries, which require decomposing a high-level objective into multiple interdependent API calls with correct planning and execution. Current approaches mainly rely on zero-shot evaluation due to the absence of training data. While proprietary closed-source models such as GPT-4 demonstrate strong reasoning abilities, smaller open-source models struggle to perform complex tool use effectively. Thus, we propose a novel training framework GOAT, which enables fine-tuning of LLM agents in a human annotation-free setting. GOAT automatically constructs synthetic datasets of goal-oriented API execution tasks directly from given API documents, equipping models with the ability to reason over interdependent calls and generate coherent responses. Through extensive experiments, we show that GOAT-trained agents achieve state-of-the-art performance across multiple existing goal-oriented benchmarks. In addition, we introduce GOATBench, a new goal-oriented API execution benchmark, and demonstrate that agents trained with GOAT also excel in this setting. These results highlight GOAT as a practical path toward building robust open-source LLM agents capable of complex reasoning and tool use.
Abstract (translated)
近期，大型语言模型（LLMs）已从传统的文本生成扩展到能够根据用户意图使用外部工具的交互式代理。然而，当前的LLM代理在处理以目标为导向的查询时仍表现出有限的能力，这些查询需要将高层次的目标分解为多个相互依赖的API调用，并进行正确的计划和执行。目前的方法主要依靠零样本评估来完成这项任务，因为缺乏训练数据。虽然像GPT-4这样的专有闭源模型展示了强大的推理能力，但较小的开源模型在有效使用复杂工具方面仍面临困难。因此，我们提出了一种新的培训框架GOAT（Goal-Oriented API Task），该框架能够在无需人工注释的情况下对LLM代理进行微调。GOAT能够从给定的API文档中自动构建目标导向型API执行任务的合成数据集，从而使模型具备处理相互依赖调用和生成连贯响应的能力。 通过广泛的实验，我们展示了使用GOAT训练的代理在多个现有的以目标为导向的基准测试中实现了最先进的性能。此外，我们还引入了GOATBench，这是一个新的以目标为导向的API执行基准，并证明了使用GOAT训练的代理在这个设置下同样表现出色。这些结果强调了GOAT是构建具有复杂推理和工具使用能力的强大开源LLM代理的实际路径。
URL
https://arxiv.org/abs/2510.12218
PDF
-
Precise Attribute Intensity Control in Large Language Models via Targeted Representation Editing
2025-10-14 03:50:22Rongzhi Zhang, Liqin Ye, Yuzhao Heng, Xiang Chen, Tong Yu, Lingkai Kong, Sudheer Chava, Chao ZhangAbstract
Precise attribute intensity control--generating Large Language Model (LLM) outputs with specific, user-defined attribute intensities--is crucial for AI systems adaptable to diverse user expectations. Current LLM alignment methods, however, typically provide only directional or open-ended guidance, failing to reliably achieve exact attribute intensities. We address this limitation with three key designs: (1) reformulating precise attribute intensity control as a target-reaching problem, rather than simple maximization; (2) training a lightweight value function via temporal-difference learning to predict final attribute intensity scores from partial generations, thereby steering LLM outputs; and (3) employing gradient-based interventions on hidden representations to navigate the model precisely towards specific attribute intensity targets. Our method enables fine-grained, continuous control over attribute intensities, moving beyond simple directional alignment. Experiments on LLaMA-3.2-3b and Phi-4-mini confirm our method's ability to steer text generation to user-specified attribute intensities with high accuracy. Finally, we demonstrate efficiency enhancements across three downstream tasks: preference data synthesis, Pareto frontier approximation and optimization, and distillation of aligned behaviors for intervention-free inference. Our code is available on this https URL
Abstract (translated)
精确的属性强度控制——即根据用户定义的具体强度生成大型语言模型（LLM）输出——对于适应多样化用户期望的人工智能系统至关重要。然而，当前的语言模型对齐方法通常只能提供方向性或开放式的指导，并不能可靠地实现确切的属性强度。为了解决这一局限性，我们提出了三个关键设计：(1) 将精确属性强度控制重新定义为目标导向问题而非简单最大化；(2) 通过时差学习训练一个轻量级的价值函数来预测从部分生成中得到的最终属性强度分数，并据此引导LLM输出；以及 (3) 使用基于梯度的方法干预隐藏表示，以使模型准确地导航到特定的属性强度目标。我们的方法能够实现细粒度、连续控制的属性强度调整，超越了简单的方向对齐。在LLaMA-3.2-3b和Phi-4-mini上的实验确认了我们的方法能够将文本生成引导至用户指定的具体属性强度上，并且精度很高。最后，我们展示了该方法在三个下游任务中的效率提升：偏好数据合成、帕累托前沿逼近与优化以及无干预推理的对齐行为精炼。我们的代码可在[此链接](https://这个位置填写实际的URL)获取。
URL
https://arxiv.org/abs/2510.12121
PDF
-
PAGE: Prompt Augmentation for text Generation Enhancement
2025-10-13 21:31:04Mauro Jose Pacchiotti, Luciana Ballejos, Mariel AleAbstract
In recent years, natural language generative models have shown outstanding performance in text generation tasks. However, when facing specific tasks or particular requirements, they may exhibit poor performance or require adjustments that demand large amounts of additional data. This work introduces PAGE (Prompt Augmentation for text Generation Enhancement), a framework designed to assist these models through the use of simple auxiliary modules. These modules, lightweight models such as classifiers or extractors, provide inferences from the input text. The output of these auxiliaries is then used to construct an enriched input that improves the quality and controllability of the generation. Unlike other generation-assistance approaches, PAGE does not require auxiliary generative models; instead, it proposes a simpler, modular architecture that is easy to adapt to different tasks. This paper presents the proposal, its components and architecture, and reports a proof of concept in the domain of requirements engineering, where an auxiliary module with a classifier is used to improve the quality of software requirements generation.
Abstract (translated)
近年来，自然语言生成模型在文本生成任务中表现出色。然而，在面对特定任务或具体需求时，这些模型可能会表现不佳或需要大量额外数据进行调整以改进性能。这项工作引入了PAGE（Prompt Augmentation for text Generation Enhancement），这是一个通过使用简单辅助模块来帮助这些模型的框架。这些轻量级模型（如分类器或提取器）会从输入文本中提供推断结果。然后，将这些辅助模块的输出用于构建增强后的输入，以提高生成的质量和可控性。与其它生成辅助方法不同的是，PAGE不需要额外的生成模型；相反，它提出了一种简单、模块化的架构，易于适应不同的任务需求。 本文介绍了该框架的提案及其组件和架构，并在软件需求工程领域展示了概念验证案例，在此案例中，使用了一个带有分类器的辅助模块来提高软件需求生成的质量。
URL
https://arxiv.org/abs/2510.13880
PDF
-
GRAVITY: A Framework for Personalized Text Generation via Profile-Grounded Synthetic Preferences
2025-10-13 21:23:46Priyanka Dey, Daniele Rosa, Wenqing Zheng, Daniel Barcklow, Jieyu Zhao, Emilio FerraraAbstract
Personalization in LLMs often relies on costly human feedback or interaction logs, limiting scalability and neglecting deeper user attributes. To reduce the reliance on human annotations, we introduce GRAVITY (Generative Response with Aligned Values, Interests, and Traits of You), a framework for generating synthetic, profile-grounded preference data that captures users' interests, values, beliefs, and personality traits. By integrating demographic, cultural, and psychological frameworks -- including Hofstede's cultural dimensions, Schwartz's basic values, the World Values Survey, and Big Five OCEAN traits -- GRAVITY synthesizes preference pairs to guide personalized content generation. We evaluate GRAVITY on book descriptions for 400 Amazon users, comparing it to prompt-based conditioning, standard fine-tuning, and naive synthetic pair generation. Profile-grounded synthetic data consistently improves generation, especially across multiple cultures (USA, Brazil, Japan, India), achieving over 4% higher preference gains across baselines, with user studies showing that GRAVITY outputs are preferred over 86% of the time. Our results show that scenario-grounded synthetic data can capture richer user variation, reduce reliance on costly annotation, and produce more engaging, user-centered content, offering a scalable path for LLM personalization.
Abstract (translated)
在大型语言模型（LLMs）中，个性化通常依赖于昂贵的人类反馈或交互日志，这限制了其可扩展性，并忽视了更深层次的用户属性。为了减少对人工注释的依赖，我们引入了一种名为GRAVITY（与您的价值观、兴趣和特质相匹配的生成式响应）的框架，用于生成基于个人资料的合成偏好数据，该数据能够捕捉用户的兴趣、价值观念、信仰和个人特质。通过整合人口统计学、文化和心理框架——包括霍夫斯泰德的文化维度理论、施瓦茨的基本价值观、世界价值观调查以及五大人格特质（开放性、尽责性、外向性、宜人性和神经质）——GRAVITY综合生成偏好对，以指导个性化内容的创作。我们在400名亚马逊用户中的书籍描述上评估了GRAVITY，并将其与基于提示的方法调优、标准微调以及简单的合成对生成方法进行了比较。结果表明，基于个人资料的合成数据在提升生成效果方面表现突出，尤其是在多文化背景（美国、巴西、日本和印度）下，相对于基准线，偏好增益提高了超过4%；用户研究显示，在86%的时间内，GRAVITY生成的内容更受欢迎。我们的结果显示，场景相关的合成数据能够捕捉到更为丰富的用户差异性，减少对昂贵注释的依赖，并且能产出更多引人入胜、以用户为中心的内容，为大型语言模型的个性化提供了一条可扩展的道路。
URL
https://arxiv.org/abs/2510.11952
PDF
-
Secret-Protected Evolution for Differentially Private Synthetic Text Generation
2025-10-13 04:05:42Tianze Wang, Zhaoyu Chen, Jian Du, Yingtai Xiao, Linjun Zhang, Qiang YanAbstract
Text data has become extremely valuable on large language models (LLMs) and even lead to general artificial intelligence (AGI). A lot of high-quality text in the real world is private and cannot be freely used due to privacy concerns. Therefore, differentially private (DP) synthetic text generation has been proposed, aiming to produce high-utility synthetic data while protecting sensitive information. However, existing DP synthetic text generation imposes uniform guarantees that often overprotect non-sensitive content, resulting in substantial utility loss and computational overhead. Therefore, we propose Secret-Protected Evolution (SecPE), a novel framework that extends private evolution with secret-aware protection. Theoretically, we show that SecPE satisfies $(\mathrm{p}, \mathrm{r})$-secret protection, constituting a relaxation of Gaussian DP that enables tighter utility-privacy trade-offs, while also substantially reducing computational complexity relative to baseline methods. Empirically, across the OpenReview, PubMed, and Yelp benchmarks, SecPE consistently achieves lower Fréchet Inception Distance (FID) and higher downstream task accuracy than GDP-based Aug-PE baselines, while requiring less noise to attain the same level of protection. Our results highlight that secret-aware guarantees can unlock more practical and effective privacy-preserving synthetic text generation.
Abstract (translated)
文本数据在大型语言模型（LLMs）中变得极其有价值，甚至可以引领通用人工智能（AGI）的发展。然而，现实世界中的大量高质量文本由于隐私问题而无法自由使用且是私有的。因此，提出了差异隐私（DP）合成文本生成方法，旨在在保护敏感信息的同时产生高价值的合成数据。但是，现有的DP合成文本生成方法通常对非敏感内容提供过度保护，导致显著的效用损失和计算开销增加。为此，我们提出了一种新的框架——秘密保护演化（Secret-Protected Evolution, SecPE），该框架通过引入对秘密信息的认识来扩展私有演化的保护措施。 从理论上讲，我们证明SecPE满足$(\mathrm{p}, \mathrm{r})$-秘密保护标准，这是一个高斯DP的放松版本，能够实现更加紧密的效用与隐私权衡，并且相对于基线方法大幅减少了计算复杂性。在实证研究中，通过OpenReview、PubMed和Yelp等基准数据集上的实验表明，SecPE在弗雷歇 inception 距离（FID）方面始终低于基于GDP的Aug-PE基线模型，并实现了更高的下游任务准确性，同时达到了相同级别的保护所需的噪声更少。我们的研究结果强调了秘密意识保证可以在隐私保护合成文本生成中解锁更多实用且有效的解决方案。
URL
https://arxiv.org/abs/2510.10990
PDF
-
IUT-Plug: A Plug-in tool for Interleaved Image-Text Generation
2025-10-13 03:19:45Zeteng Lin, Xingxing Li, Wen You, Xiaoyang Li, Zehan Lu, Yujun Cai, Jing TangAbstract
Existing vision language models (VLMs), including GPT-4 and DALL-E, often struggle to preserve logic, object identity, and style in multimodal image-text generation. This limitation significantly hinders the generalization capability of VLMs in complex image-text input-output scenarios. To address this issue, we propose IUT-Plug, a module grounded in an Image Understanding Tree (IUT), which enhances existing interleaved VLMs through explicit structured reasoning, thereby mitigating context drift in logic, entity identity, and style. The proposed framework operates in two stages. (1) A dynamic IUT-Plug extraction module parses visual scenes into hierarchical symbolic structures. (2) A coordinated narrative-flow and image synthesis mechanism ensures cross-modal consistency. To evaluate our approach, we construct a novel benchmark based on 3,000 real human-generated question-answer pairs over fine-tuned large models, introducing a dynamic evaluation protocol for quantifying context drift in interleaved VLMs. Experimental results demonstrate that IUT-Plug not only improves accuracy on established benchmarks but also effectively alleviates the three critical forms of context drift across diverse multimodal question answering (QA) scenarios.
Abstract (translated)
现有的视觉语言模型（VLM），包括GPT-4和DALL-E，在多模态图像文本生成中通常难以保持逻辑、对象身份和风格的一致性。这一局限极大地阻碍了VLM在复杂图像文字输入输出场景中的泛化能力。为了解决这个问题，我们提出了基于图像理解树（IUT）的IUT-Plug模块，通过显式的结构化推理来增强现有的交错式视觉语言模型（interleaved VLM），从而减轻逻辑、实体身份和风格方面的上下文漂移问题。 该框架分为两个阶段操作： 1. 动态IUT-Plug提取模块将视觉场景解析为分层的符号结构。 2. 协同叙事流程与图像合成机制确保跨模态的一致性。 为了评估我们的方法，我们构建了一个基于3,000对真实的人类生成的问题答案对的新基准测试，并引入了一种量化交错式VLM中上下文漂移的动态评价协议。实验结果表明，IUT-Plug不仅在现有的基准上提高了准确性，而且还有效地缓解了跨多种多模态问答场景中的三种关键形式的上下文漂移问题。
URL
https://arxiv.org/abs/2510.10969
PDF
-
BitMar: Low-Bit Multimodal Fusion with Episodic Memory for Edge Devices
2025-10-12 11:59:41Euhid Aman, Esteban Carlin, Hsing-Kuo Pao, Giovanni Beltrame, Ghaluh Indah Permata Sari, Yie-Tarng ChenAbstract
Cross-attention transformers and other multimodal vision-language models excel at grounding and generation; however, their extensive, full-precision backbones make it challenging to deploy them on edge devices. Memory-augmented architectures enhance the utilization of past context; however, most works rarely pair them with aggressive edge-oriented quantization. We introduce BitMar, a quantized multimodal transformer that proposes an external human-like episodic memory for effective image-text generation on hardware with limited resources. BitMar utilizes 1.58-bit encoders, one for text (BitNet-style) and one for vision (DiNOv2-based), to create compact embeddings that are combined and used to query a fixed-size key-value episodic memory. During vector retrieval, the BitNet decoder applies per-layer conditioning, which increases the contextual relevance of generated content. The decoder also employs attention sinks with a sliding-window mechanism to process long or streaming inputs under tight memory budgets. The combination of per-layer conditioning and sliding-window attention achieves a strong quality-speed trade-off, delivering competitive captioning and multimodal understanding at low latency with a small model footprint. These characteristics make BitMar well-suited for edge deployment.
Abstract (translated)
跨模态视觉-语言模型（如带有交叉注意力的变压器）在图像文本生成和理解方面表现出色，但其庞大的全精度骨干网络使得它们难以部署到资源受限的边缘设备上。尽管内存增强架构提高了对过去上下文的利用效率，但大多数研究很少将这些架构与激进的面向边缘设备的量化技术相结合。 我们引入了BitMar——一种采用人类类比的外部分布式记忆系统的量化多模态变压器模型。这种设计使得在资源有限的硬件上有效进行图像-文本生成成为可能。BitMar采用了1.58位编码器，分别用于处理文本（基于BitNet风格）和视觉信息（基于DiNOv2），从而创建紧凑的嵌入式表示，并将其与固定大小的关键值外部记忆结合使用以查询相关信息。 在向量检索过程中，BitNet解码器应用逐层条件调节技术，这能显著提高生成内容的相关性。此外，解码器还利用带有滑动窗口机制的关注力汇节点来处理长输入或流式数据，在严格的内存预算限制下进行高效运算。这种结合了逐层条件调节和滑动窗口注意力的方法在保证高质量输出的同时实现了高效的计算速度。 这些特性使得BitMar非常适合边缘设备部署，能够在低延迟条件下实现竞争性的描述生成与多模态理解，并且模型尺寸较小。
URL
https://arxiv.org/abs/2510.10560
PDF
-
RAG-IGBench: Innovative Evaluation for RAG-based Interleaved Generation in Open-domain Question Answering
2025-10-11 03:06:39Rongyang Zhang, Yuqing Huang, Chengqiang Lu, Qimeng Wang, Yan Gao, Yi Wu, Yao Hu, Yin Xu, Wei Wang, Hao Wang, Enhong ChenAbstract
In real-world scenarios, providing user queries with visually enhanced responses can considerably benefit understanding and memory, underscoring the great value of interleaved image-text generation. Despite recent progress, like the visual autoregressive model that unifies text and image processing in a single transformer architecture, generating high-quality interleaved content remains challenging. Moreover, evaluations of these interleaved sequences largely remain underexplored, with existing benchmarks often limited by unimodal metrics that inadequately assess the intricacies of combined image-text outputs. To address these issues, we present RAG-IGBench, a thorough benchmark designed specifically to evaluate the task of Interleaved Generation based on Retrieval-Augmented Generation (RAG-IG) in open-domain question answering. RAG-IG integrates multimodal large language models (MLLMs) with retrieval mechanisms, enabling the models to access external image-text information for generating coherent multimodal content. Distinct from previous datasets, RAG-IGBench draws on the latest publicly available content from social platforms and introduces innovative evaluation metrics that measure the quality of text and images, as well as their consistency. Through extensive experiments with state-of-the-art MLLMs (both open-source and proprietary) on RAG-IGBench, we provide an in-depth analysis examining the capabilities and limitations of these models. Additionally, we validate our evaluation metrics by demonstrating their high correlation with human assessments. Models fine-tuned on RAG-IGBench's training set exhibit improved performance across multiple benchmarks, confirming both the quality and practical utility of our dataset. Our benchmark is available at this https URL.
Abstract (translated)
在实际应用场景中，通过视觉增强的响应来提供用户查询可以显著提高理解和记忆的效果，这突显了交错式图像-文本生成的巨大价值。尽管最近取得了进展，例如将文字和图像处理统一在一个Transformer架构中的视觉自回归模型，但要生成高质量的交错式内容仍然具有挑战性。此外，对这些交错序列的评估仍很大程度上被忽略，现有的基准测试往往受限于单模态度量标准，无法充分评价组合式图像-文本输出的复杂性。 为解决这些问题，我们提出了RAG-IGBench，这是一个专门针对基于检索增强生成（Retrieval-Augmented Generation, RAG）任务进行交错式生成评估的全面基准。RAG-IG结合了多模态大型语言模型（Multimodal Large Language Models, MLLMs）与检索机制，使这些模型能够访问外部图像文本信息以生成连贯的多模态内容。不同于之前的语料库，RAG-IGBench借鉴了最新的公开社交平台上的内容，并引入了一系列创新评估指标来衡量文本和图片的质量及其一致性。 通过对最先进的MLLMs（包括开源与专有）进行广泛的实验测试，并在RAG-IGBench上提供深入分析，我们揭示了这些模型的能力及局限性。此外，通过证明我们的评价标准与人工判断的高度相关性，进一步验证了我们的评估指标的有效性。基于RAG-IGBench训练集微调的模型，在多个基准测试中的表现都有所提升，这证实了我们数据集的质量和实用性。 我们的基准可以在以下网址获得：[请在此处插入链接]。
URL
https://arxiv.org/abs/2512.05119
PDF
-
FLRC: Fine-grained Low-Rank Compressor for Efficient LLM Inference
2025-10-10 12:35:09Yu-Chen Lu, Chong-Yan Chen, Chi-Chih Chang, Yu-Fang Hu, Kai-Chiang WuAbstract
Although large language models (LLM) have achieved remarkable performance, their enormous parameter counts hinder deployment on resource-constrained hardware. Low-rank compression can reduce both memory usage and computational demand, but applying a uniform compression ratio across all layers often leads to significant performance degradation, and previous methods perform poorly during decoding. To address these issues, we propose the Fine-grained Low-Rank Compressor (FLRC), which efficiently determines an optimal rank allocation for each layer, and incorporates progressive low-rank decoding to maintain text generation quality. Comprehensive experiments on diverse benchmarks demonstrate the superiority of FLRC, achieving up to a 17% improvement in ROUGE-L on summarization tasks compared to state-of-the-art low-rank compression methods, establishing a more robust and efficient framework to improve LLM inference.
Abstract (translated)
尽管大型语言模型（LLM）已经取得了显著的性能成果，但其庞大的参数数量阻碍了它们在资源受限硬件上的部署。低秩压缩可以减少内存使用和计算需求，但是对所有层应用统一的压缩比率通常会导致明显的性能下降，并且之前的方法在解码过程中表现不佳。为了解决这些问题，我们提出了细粒度低秩压缩器（FLRC），它可以高效地为每一层确定最优的秩分配，并结合渐进式低秩解码来保持文本生成的质量。在各种基准测试中的全面实验表明了FLRC的优越性，在摘要任务中与最先进的低秩压缩方法相比，ROUGE-L评分提高了最多17%，从而建立了更为稳健和高效的框架以改进LLM的推理能力。
URL
https://arxiv.org/abs/2510.09332
PDF
-
A Unified Biomedical Named Entity Recognition Framework with Large Language Models
2025-10-10 01:33:54Tengxiao Lv, Ling Luo, Juntao Li, Yanhua Wang, Yuchen Pan, Chao Liu, Yanan Wang, Yan Jiang, Huiyi Lv, Yuanyuan Sun, Jian Wang, Hongfei LinAbstract
Accurate recognition of biomedical named entities is critical for medical information extraction and knowledge discovery. However, existing methods often struggle with nested entities, entity boundary ambiguity, and cross-lingual generalization. In this paper, we propose a unified Biomedical Named Entity Recognition (BioNER) framework based on Large Language Models (LLMs). We first reformulate BioNER as a text generation task and design a symbolic tagging strategy to jointly handle both flat and nested entities with explicit boundary annotation. To enhance multilingual and multi-task generalization, we perform bilingual joint fine-tuning across multiple Chinese and English datasets. Additionally, we introduce a contrastive learning-based entity selector that filters incorrect or spurious predictions by leveraging boundary-sensitive positive and negative samples. Experimental results on four benchmark datasets and two unseen corpora show that our method achieves state-of-the-art performance and robust zero-shot generalization across languages. The source codes are freely available at this https URL.
Abstract (translated)
生物医学命名实体的准确识别对于医疗信息提取和知识发现至关重要。然而，现有方法在处理嵌套实体、边界模糊以及跨语言泛化方面常常遇到困难。本文提出了一种基于大型语言模型（LLMs）的统一生物医学命名实体识别（BioNER）框架。首先，我们将BioNER重新定义为一个文本生成任务，并设计了一种符号标记策略，以明确标注边界的双重方式处理扁平和嵌套实体。为了增强多语言和多任务泛化能力，我们在多个中文和英文数据集上进行了双语联合微调。此外，我们引入了一个基于对比学习的实体选择器，通过利用边界敏感的正负样本来过滤错误或虚假预测。 在四个基准数据集和两个未见过的数据集中进行的实验结果表明，我们的方法实现了最先进的性能，并且具有跨语言的稳健零样本泛化能力。源代码可在此URL免费获取。
URL
https://arxiv.org/abs/2510.08902
PDF
-
AI-Driven Radiology Report Generation for Traumatic Brain Injuries
2025-10-09 17:39:04Riadh Bouslimi, Houda Trabelsi, Wahiba Ben Abdssalem Karaa, Hana HedhliAbstract
Traumatic brain injuries present significant diagnostic challenges in emergency medicine, where the timely interpretation of medical images is crucial for patient outcomes. In this paper, we propose a novel AI-based approach for automatic radiology report generation tailored to cranial trauma cases. Our model integrates an AC-BiFPN with a Transformer architecture to capture and process complex medical imaging data such as CT and MRI scans. The AC-BiFPN extracts multi-scale features, enabling the detection of intricate anomalies like intracranial hemorrhages, while the Transformer generates coherent, contextually relevant diagnostic reports by modeling long-range dependencies. We evaluate the performance of our model on the RSNA Intracranial Hemorrhage Detection dataset, where it outperforms traditional CNN-based models in both diagnostic accuracy and report generation. This solution not only supports radiologists in high-pressure environments but also provides a powerful educational tool for trainee physicians, offering real-time feedback and enhancing their learning experience. Our findings demonstrate the potential of combining advanced feature extraction with transformer-based text generation to improve clinical decision-making in the diagnosis of traumatic brain injuries.
Abstract (translated)
颅脑损伤在急诊医学中诊断难度大，及时解读医疗影像对于患者的预后至关重要。本文提出了一种基于人工智能的新型自动放射学报告生成方法，专门针对颅骨创伤病例设计。我们的模型结合了AC-BiFPN和Transformer架构，以捕获并处理复杂的医学成像数据，如CT和MRI扫描。其中，AC-BiFPN提取多尺度特征，能够检测到诸如脑内出血等细微异常情况；而Transformer通过建模长程依赖关系生成连贯且上下文相关的诊断报告。 我们在RSNA脑内出血检测数据集上评估了模型的性能，结果显示我们的方法在诊断准确性和报告生成方面都优于传统的基于CNN（卷积神经网络）的模型。这一解决方案不仅支持放射科医生在高压环境中工作，还为实习医师提供了一种强大的教育工具，能够实时反馈并提高其学习体验。 我们的研究结果展示了将高级特征提取与基于Transformer的文本生成相结合以改善颅脑损伤诊断中临床决策制定的潜力。
URL
https://arxiv.org/abs/2510.08498
PDF
-
Lossless Vocabulary Reduction for Auto-Regressive Language Models
2025-10-09 11:38:48Daiki Chijiwa, Taku Hasegawa, Kyosuke Nishida, Shin'ya Yamaguchi, Tomoya Ohba, Tamao Sakao, Susumu TakeuchiAbstract
Tokenization -- the process of decomposing a given text into a sequence of subwords called tokens -- is one of the key components in the development of language models. Particularly, auto-regressive language models generate texts token by token, i.e., by predicting the next-token distribution given the previous ones, and thus tokenization directly affects their efficiency in text generation. Since each language model has their own vocabulary as a set of possible tokens, they struggle to cooperate with each other at the level of next-token distributions such as model ensemble. In this paper, we establish a theoretical framework of lossless vocabulary reduction, which efficiently converts a given auto-regressive language model into the one with an arbitrarily small vocabulary without any loss in accuracy. As an application, we demonstrate that language models with different tokenization can cooperate with each other efficiently through their maximal common vocabulary.
Abstract (translated)
令牌化——即将给定的文本分解为一系列称为标记的子词的过程——是语言模型开发的关键组成部分。特别是，自回归语言模型通过逐个标记生成文本，即根据之前的标记预测下一个标记的概率分布，因此，令牌化直接影响它们在文本生成中的效率。由于每个语言模型都有其独特的词汇集作为可能的标记集合，它们难以在其概率分布层面（如模型集成）上相互合作。本文中，我们建立了一个无损词汇缩减的理论框架，该框架可以高效地将给定自回归语言模型转换为具有任意小词汇量的形式，而不会牺牲准确性。作为一种应用，我们展示了不同令牌化方法的语言模型可以通过它们的最大公共词汇库有效地进行协作。 这段翻译解释了令牌化的概念和其在语言模型中的重要性，并概述了一种新的理论框架，该框架能够将自回归语言模型转换为具有更小词汇量的形式而不损失精度。此外，还描述了一个具体应用案例，即展示不同令牌化方法的语言模型如何通过共享的最大公共词汇进行协作。
URL
https://arxiv.org/abs/2510.08102
PDF
-
Comprehensiveness Metrics for Automatic Evaluation of Factual Recall in Text Generation
2025-10-09 08:22:24Adam Dejl, James Barry, Alessandra Pascale, Javier Carnerero CanoAbstract
Despite demonstrating remarkable performance across a wide range of tasks, large language models (LLMs) have also been found to frequently produce outputs that are incomplete or selectively omit key information. In sensitive domains, such omissions can result in significant harm comparable to that posed by factual inaccuracies, including hallucinations. In this study, we address the challenge of evaluating the comprehensiveness of LLM-generated texts, focusing on the detection of missing information or underrepresented viewpoints. We investigate three automated evaluation strategies: (1) an NLI-based method that decomposes texts into atomic statements and uses natural language inference (NLI) to identify missing links, (2) a Q&A-based approach that extracts question-answer pairs and compares responses across sources, and (3) an end-to-end method that directly identifies missing content using LLMs. Our experiments demonstrate the surprising effectiveness of the simple end-to-end approach compared to more complex methods, though at the cost of reduced robustness, interpretability and result granularity. We further assess the comprehensiveness of responses from several popular open-weight LLMs when answering user queries based on multiple sources.
Abstract (translated)
尽管大型语言模型（LLM）在广泛的任务中表现出色，但它们也被发现经常会生成不完整或有选择性地省略关键信息的输出。在这种情况下，在敏感领域，这些遗漏可能会造成与事实错误相当的危害，包括幻觉现象。在这项研究中，我们解决了一个评估LLM生成文本全面性的挑战，重点在于检测缺失的信息或被边缘化的观点。我们调查了三种自动评估策略：（1）基于自然语言推理（NLI）的方法，将文本分解为原子陈述，并使用自然语言推理来识别缺失的联系；（2）基于问答的方式，提取问题-答案对并跨源比较响应；以及（3）一种端到端方法，直接利用LLM来识别缺少的内容。我们的实验表明，尽管简单端到端的方法在稳健性、可解释性和结果粒度方面有所牺牲，但它比复杂方法更有效地检测到了缺失信息。我们进一步评估了几个流行的开源LLM在回答基于多个来源的用户查询时响应的全面性。
URL
https://arxiv.org/abs/2510.07926
PDF
