Text_Generation
-
LLM-Based Intelligent Notification Composition: From Static Personalization to Context-Aware Persuasive Messaging
Abstract
Push notifications remain among the most direct channels through which digital platforms engage users, yet existing approaches have invested heavily in who to notify, when to notify, and what to recommend, while leaving how to communicate as the least-optimized stage. This paper argues that message quality is an independent, underinvested lever, and that LLMs create their most differentiated value precisely at this layer. We make three contributions. First, we define notification message quality along six dimensions (contextual relevance, clarity, actionability, novelty handling, linguistic freshness, and persuasive appropriateness) and show how LLM-based composition improves each relative to templates. Across reviewed deployments, reported improvements range from +8% to +14.5% CTR over static templates and +1% to +2.5% over mature slot-filling systems, though these span heterogeneous systems and should not be treated as directly comparable. Second, we provide an architectural attribution analysis disentangling message generation from adjacent components (targeting, ranking, timing), arguing that observed gains are frequently misattributed to text generation alone. Third, we introduce a three-criterion decision framework specifying when LLM generation is and is not the binding constraint. We support these arguments through a PRISMA-guided survey (28 sources from 142 screened), examine domain-specific applications across social media, food delivery, and e-commerce, and propose a unified architectural framework with budget-aware routing, grounded generation, candidate ranking, diversity controls, and online learning.
Abstract (translated)
URL
https://arxiv.org/abs/2605.16264
PDF
-
MzansiText and MzansiLM: An Open Corpus and Decoder-Only Language Model for South African Languages
2026-03-21 09:35:40Anri Lombard, Simbarashe Mawere, Temi Aina, Ethan Wolff, Sbonelo Gumede, Elan Novick, Francois Meyer, Jan BuysAbstract
Decoder-only language models can be adapted to diverse tasks through instruction finetuning, but the extent to which this generalizes at small scale for low-resource languages remains unclear. We focus on the languages of South Africa, where we are not aware of a publicly available decoder-only model that explicitly targets all eleven official written languages, nine of which are low-resource. We introduce MzansiText, a curated multilingual pretraining corpus with a reproducible filtering pipeline, and MzansiLM, a 125M-parameter language model trained from scratch. We evaluate MzansiLM on natural language understanding and generation using three adaptation regimes: monolingual task-specific finetuning, multilingual task-specific finetuning, and general multi-task instruction finetuning. Monolingual task-specific finetuning achieves strong performance on data-to-text generation, reaching 20.65 BLEU on isiXhosa and competing with encoder-decoder baselines over ten times larger. Multilingual task-specific finetuning benefits closely related languages on topic classification, achieving 78.5% macro-F1 on isiXhosa news classification. While MzansiLM adapts effectively to supervised NLU and NLG tasks, few-shot reasoning remains challenging at this model size, with performance near chance even for much larger decoder-only models. We release MzansiText and MzansiLM to provide a reproducible decoder-only baseline and clear guidance on adaptation strategies for South African languages at small scale.
Abstract (translated)
URL
https://arxiv.org/abs/2603.20732
PDF
-
Cross-modal Fuzzy Alignment Network for Text-Aerial Person Retrieval and A Large-scale Benchmark
2026-03-21 09:02:44Yifei Deng, Chenglong Li, Yuyang Zhang, Guyue Hu, Jin TangAbstract
Text-aerial person retrieval aims to identify targets in UAV-captured images from eyewitness descriptions, supporting intelligent transportation and public security applications. Compared to ground-view text--image person retrieval, UAV-captured images often suffer from degraded visual information due to drastic variations in viewing angles and flight altitudes, making semantic alignment with textual descriptions very challenging. To address this issue, we propose a novel Cross-modal Fuzzy Alignment Network, which quantifies the token-level reliability by fuzzy logic to achieve accurate fine-grained alignment and incorporates ground-view images as a bridge agent to further mitigate the gap between aerial images and text descriptions, for text--aerial person retrieval. In particular, we design the Fuzzy Token Alignment module that employs the fuzzy membership function to dynamically model token-level association strength and suppress the influence of unobservable or noisy tokens. It can alleviate the semantic inconsistencies caused by missing visual cues and significantly enhance the robustness of token-level semantic alignment. Moreover, to further mitigate the gap between aerial images and text descriptions, we design a Context-Aware Dynamic Alignment module to incorporate the ground-view agent as a bridge in text--aerial alignment and adaptively combine direct alignment and agent-assisted alignment to improve the robustness. In addition, we construct a large-scale benchmark dataset called AERI-PEDES by using a chain-of-thought to decompose text generation into attribute parsing, initial captioning, and refinement, thus boosting textual accuracy and semantic consistency. Experiments on AERI-PEDES and TBAPR demonstrate the superiority of our method.
Abstract (translated)
URL
https://arxiv.org/abs/2603.20721
PDF
-
Diffutron: A Masked Diffusion Language Model for Turkish Language
2026-03-20 19:52:37\c{S}uayp Talha Kocabay, Talha R\"uzgar Akku\c{s}Abstract
Masked Diffusion Language Models (MDLMs) have emerged as a compelling non-autoregressive alternative to standard large language models; however, their application to morphologically rich languages remains limited. In this paper, we introduce $\textit{Diffutron}$, a masked diffusion language model specifically designed for Turkish. Our approach leverages a resource-efficient training pipeline, starting with LoRA-based continual pre-training of a multilingual encoder on a large-scale corpus. To enable generative capabilities, we employ a progressive instruction-tuning strategy, sequentially adapting the model on general and task-specific instruction sets. Experimental results across comprehensive benchmarks demonstrate that, despite its compact size, our model achieves competitive performance compared to existing multi-billion-parameter baselines. These findings validate the effectiveness of masked diffusion modeling combined with multi-stage tuning for non-autoregressive text generation in Turkish.
Abstract (translated)
掩码扩散语言模型（MDLMs）已成为标准大语言模型的一个引人注目的非自回归替代方案；然而，其在形态丰富语言上的应用仍较为有限。本文中，我们介绍了专为土耳其语设计的掩码扩散语言模型$\textit{Diffutron}$。我们的方法采用资源高效的训练流程，首先基于LoRA对多语言编码器在大规模语料上进行持续预训练。为赋予模型生成能力，我们采用渐进式指令微调策略，依次在通用和任务特定指令集上对模型进行适配。综合基准测试的实验结果表明，尽管模型规模较小，我们的方法仍能达到与现有数十亿参数基线相比具有竞争力的性能。这些发现验证了掩码扩散建模结合多阶段微调在土耳其语非自回归文本生成中的有效性。
URL
https://arxiv.org/abs/2603.20466
PDF
-
D5P4: Partition Determinantal Point Process for Diversity in Parallel Discrete Diffusion Decoding
2026-03-19 17:02:08Jonathan Lys, Vincent Gripon, Bastien Pasdeloup, Axel Marmoret, Lukas Mauch, Fabien Cardinaux, Ghouthi Boukli HaceneAbstract
Discrete diffusion models are promising alternatives to autoregressive approaches for text generation, yet their decoding methods remain under-studied. Standard decoding methods for autoregressive models, such as beam search, do not directly apply to iterative denoising, and existing diffusion decoding techniques provide limited control over in-batch diversity. To bridge this gap, we introduce a generalized beam-search framework for discrete diffusion that generates candidates in parallel and supports modular beam-selection objectives. As a diversity-focused instantiation, we propose D5P4, which formulates the selection step as MAP inference over a Determinantal Point Process. Leveraging a scalable greedy solver, D5P4 maintains multi-GPU compatibility and enables an explicit trade-off between model probability and target diversity with near-zero compute overhead. Experiments on free-form generation and question answering demonstrate that D5P4 improves diversity over strong baselines while maintaining competitive generation quality.
Abstract (translated)
URL
https://arxiv.org/abs/2603.19146
PDF
-
Secure Linear Alignment of Large Language Models
2026-03-19 13:43:32Matt Gorbett, Suman JanaAbstract
Language models increasingly appear to learn similar representations, despite differences in training objectives, architectures, and data modalities. This emerging compatibility between independently trained models introduces new opportunities for cross-model alignment to downstream objectives. Moreover, it unlocks new potential application domains, such as settings where security, privacy, or competitive constraints prohibit direct data or model sharing. In this work, we propose a privacy-preserving framework that exploits representational convergence to enable cross-silo inference between independent language models. The framework learns an affine transformation over a shared public dataset and applies homomorphic encryption to protect client queries during inference. By encrypting only the linear alignment and classification operations, the method achieves sub-second inference latency while maintaining strong security guarantees. We support this framework with an empirical investigation into representational convergence, in which we learn linear transformations between the final hidden states of independent models. We evaluate these cross-model mappings on embedding classification and out-of-distribution detection, observing minimal performance degradation across model pairs. Additionally, we show for the first time that linear alignment sometimes enables text generation across independently trained models.
Abstract (translated)
URL
https://arxiv.org/abs/2603.18908
PDF
-
A Human-in/on-the-Loop Framework for Accessible Text Generation
2026-03-19 13:23:34Lourdes Moreno, Paloma Mart\'inezAbstract
Plain Language and Easy-to-Read formats in text simplification are essential for cognitive accessibility. Yet current automatic simplification and evaluation pipelines remain largely automated, metric-driven, and fail to reflect user comprehension or normative standards. This paper introduces a hybrid framework that explicitly integrates human participation into LLM-based accessible text generation. Human-in-the-Loop (HiTL) contributions guide adjustments during generation, while Human-on-the-Loop (HoTL) supervision ensures systematic post-generation review. Empirical evidence from user studies and annotated resources is operationalized into (i) checklists aligned with standards, (ii) Event-Condition-Action trigger rules for activating expert oversight, and (iii) accessibility Key Performance Indicators (KPIs). The framework shows how human-centered mechanisms can be encoded for evaluation and reused to provide structured feedback that improves model adaptation. By embedding the human role in both generation and supervision, it establishes a traceable, reproducible, and auditable process for creating and evaluating accessible texts. In doing so, it integrates explainability and ethical accountability as core design principles, contributing to more transparent and inclusive NLP systems.
Abstract (translated)
Plain Language和Easy-to-Read格式在文本简化中对认知可访问性至关重要。然而，当前的自动简化与评估流程仍高度自动化、指标驱动，未能反映用户理解程度或规范标准。本文提出一个混合框架，将人类参与明确整合到基于大语言模型的可访问文本生成过程中。Human-in-the-Loop（HiTL）贡献在生成过程中指导调整，而Human-on-the-Loop（HoTL）监督则确保系统化的生成后审查。来自用户研究和标注资源的实证证据被转化为：（i）与标准对齐的检查清单；（ii）用于激活专家监督的事件-条件-动作触发规则；（iii）可访问性关键绩效指标（KPIs）。该框架展示了如何将人本机制编码用于评估，并复用其提供结构化反馈以改进模型适应。通过将人类角色嵌入生成与监督双环节，它建立了可追溯、可复现、可审计的可访问文本创建与评估流程。在此过程中，框架将可解释性与伦理问责作为核心设计原则，推动构建更透明、更具包容性的自然语言处理系统。
URL
https://arxiv.org/abs/2603.18879
PDF
-
Mi:dm K 2.5 Pro
2026-03-19 11:37:06KT Tech innovation GroupAbstract
The evolving LLM landscape requires capabilities beyond simple text generation, prioritizing multi-step reasoning, long-context understanding, and agentic workflows. This shift challenges existing models in enterprise environments, especially in Korean-language and domain-specific scenarios where scaling is insufficient. We introduce Mi:dm K 2.5 Pro, a 32B parameter flagship LLM designed to address enterprise-grade complexity through reasoning-focused optimization. Our methodology builds a robust data foundation via a quality-centric curation pipeline utilizing abstract syntax tree (AST) analysis for code, gap-filling synthesis for mathematics, and an LLM-based quality evaluator. Pre-training scales the model via layer-predictor-based Depth Upscaling (DuS) and a progressive strategy supporting a 128K token context window. Post-training introduces a specialized multi-stage pipeline, including Reasoning SFT, model merging, and asynchronous reinforcement learning (RL), to develop complex problem-solving skills. "Fusion Training" then rebalances these capabilities with conversational fluency, consistent response styling, and reliable tool-use. The evaluations show that Mi:dm K 2.5 Pro achieves competitive performance against leading global and domestic models. In addition, it sets state-of-the-art results on Korean-specific benchmarks, showcasing deep linguistic and cultural understanding. Finally, Responsible AI evaluations validate safety against attacks, ensuring a secure profile for deployment with a balance of harmlessness and responsiveness.
Abstract (translated)
URL
https://arxiv.org/abs/2603.18788
PDF
-
The Truncation Blind Spot: How Decoding Strategies Systematically Exclude Human-Like Token Choices
2026-03-19 04:36:22Esteban Garces Arias, Nurzhan Sapargali, Christian Heumann, Matthias A{\ss}enmacherAbstract
Standard decoding strategies for text generation, including top-k, nucleus sampling, and contrastive search, select tokens based on likelihood, restricting selection to high-probability regions. Human language production operates differently: tokens are chosen for communicative appropriateness rather than statistical frequency. This mismatch creates a truncation blind spot: contextually appropriate but statistically rare tokens remain accessible to humans yet unreachable by likelihood-based decoding. We hypothesize this contributes to the detectability of machine-generated text. Analyzing over 1.8 million texts across eight language models, five decoding strategies, and 53 hyperparameter configurations, we find that 8-18% of human-selected tokens fall outside typical truncation boundaries. Simple classifiers trained on predictability and lexical diversity achieve remarkable detection rates. Crucially, neither model scale nor architecture correlates strongly with detectability; truncation parameters account for most variance. Configurations achieving low detectability often produce incoherent text, indicating that evading detection and producing natural text are distinct objectives. These findings suggest detectability is enhanced by likelihood-based token selection, not merely a matter of model capability.
Abstract (translated)
文本生成的标准解码策略（包括top-k采样、核心采样和对比搜索）均基于概率选择词元，并将选择范围限制在高概率区域。而人类语言产出的运作方式不同：词元的选择基于交际恰当性而非统计频率。这种错配造成了一个“截断盲区”：那些在语境中恰当但统计上罕见的词元，人类可以自然使用，却无法被基于概率的解码策略所触及。我们假设这加剧了机器生成文本的可检测性。通过分析涵盖八个语言模型、五种解码策略及五十三种超参数配置的超过180万条文本，我们发现8%至18%的人类选词超出了常规截断边界。基于可预测性和词汇多样性训练的简单分类器实现了显著的检测率。关键在于，模型规模或架构与可检测性并无强相关性，截断参数解释了大部分方差。那些实现低检测率的配置往往产生不连贯的文本，这表明规避检测与生成自然文本是两个截然不同的目标。这些发现表明，基于概率的词元选择增强了文本的可检测性，这远非单纯模型能力的问题。
URL
https://arxiv.org/abs/2603.18482
PDF
-
Modeling the human lexicon under temperature variations: linguistic factors, diversity and typicality in LLM word associations
2026-03-18 18:10:02Maria Andueza Rodriguez, Marie Candito, Richard HuygheAbstract
Large language models (LLMs) achieve impressive results in terms of fluency in text generation, yet the nature of their linguistic knowledge - in particular the human-likeness of their internal lexicon - remains uncertain. This study compares human and LLM-generated word associations to evaluate how accurately models capture human lexical patterns. Using English cue-response pairs from the SWOW dataset and newly generated associations from three LLMs (Mistral-7B, Llama-3.1-8B, and Qwen-2.5-32B) across multiple temperature settings, we examine (i) the influence of lexical factors such as word frequency and concreteness on cue-response pairs, and (ii) the variability and typicality of LLM responses compared to human responses. Results show that all models mirror human trends for frequency and concreteness but differ in response variability and typicality. Larger models such as Qwen tend to emulate a single "prototypical" human participant, generating highly typical but minimally variable responses, while smaller models such as Mistral and Llama produce more variable yet less typical responses. Temperature settings further influence this trade-off, with higher values increasing variability but decreasing typicality. These findings highlight both the similarities and differences between human and LLM lexicons, emphasizing the need to account for model size and temperature when probing LLM lexical representations.
Abstract (translated)
大语言模型（LLMs）在文本生成的流畅性方面取得了令人瞩目的成果，然而其语言知识的本质——尤其是其内部词汇的人类相似性——仍不明确。本研究通过比较人类与LLM生成的词汇联想，评估模型在多大程度上能捕捉人类的词汇模式。利用SWOW数据集的英语线索-反应词对，以及三个LLM（Mistral-7B、Llama-3.1-8B和Qwen-2.5-32B）在多种温度设置下新生成的联想，我们考察了（i）词汇频率和具体性等词汇因素对线索-反应词对的影响，以及（ii）LLM反应与人类反应在变异性和典型性上的差异。结果表明，所有模型在频率和具体性上均反映了人类趋势，但在反应变异性和典型性上存在差异。Qwen等较大模型倾向于模仿单一“原型”人类参与者，生成高度典型但变异极小的反应；而Mistral和Llama等较小模型则产生更多变异但典型性较低的反应。温度设置进一步影响这种权衡，较高的温度值会增加变异但降低典型性。这些发现揭示了人类与LLM词汇的异同，强调在探究LLM词汇表征时需考虑模型规模和温度设置。
URL
https://arxiv.org/abs/2603.18171
PDF
-
Adaptive Contracts for Cost-Effective AI Delegation
2026-03-17 23:31:01Eden Saig, Tamar Garbuz, Ariel D. Procaccia, Inbal Talgam-Cohen, Jamie Tucker-FoltzAbstract
When organizations delegate text generation tasks to AI providers via pay-for-performance contracts, expected payments rise when evaluation is noisy. As evaluation methods become more elaborate, the economic benefits of decreased noise are often overshadowed by increased evaluation costs. In this work, we introduce adaptive contracts for AI delegation, which allow detailed evaluation to be performed selectively after observing an initial coarse signal in order to conserve resources. We make three sets of contributions: First, we provide efficient algorithms for computing optimal adaptive contracts under natural assumptions or when core problem dimensions are small, and prove hardness of approximation in the general unstructured case. We then formulate alternative models of randomized adaptive contracts and discuss their benefits and limitations. Finally, we empirically demonstrate the benefits of adaptivity over non-adaptive baselines using question-answering and code-generation datasets.
Abstract (translated)
URL
https://arxiv.org/abs/2603.17212
PDF
-
Differential Harm Propensity in Personalized LLM Agents: The Curious Case of Mental Health Disclosure
2026-03-17 16:16:35Caglar YildirimAbstract
Large language models (LLMs) are increasingly deployed as tool-using agents, shifting safety concerns from harmful text generation to harmful task completion. Deployed systems often condition on user profiles or persistent memory, yet agent safety evaluations typically ignore personalization signals. To address this gap, we investigated how mental health disclosure, a sensitive and realistic user-context cue, affects harmful behavior in agentic settings. Building on the AgentHarm benchmark, we evaluated frontier and open-source LLMs on multi-step malicious tasks (and their benign counterparts) under controlled prompt conditions that vary user-context personalization (no bio, bio-only, bio+mental health disclosure) and include a lightweight jailbreak injection. Our results reveal that harmful task completion is non-trivial across models: frontier lab models (e.g., GPT 5.2, Claude Sonnet 4.5, Gemini 3-Pro) still complete a measurable fraction of harmful tasks, while an open model (DeepSeek 3.2) exhibits substantially higher harmful completion. Adding a bio-only context generally reduces harm scores and increases refusals. Adding an explicit mental health disclosure often shifts outcomes further in the same direction, though effects are modest and not uniformly reliable after multiple-testing correction. Importantly, the refusal increase also appears on benign tasks, indicating a safety--utility trade-off via over-refusal. Finally, jailbreak prompting sharply elevates harm relative to benign conditions and can weaken or override the protective shift induced by personalization. Taken together, our results indicate that personalization can act as a weak protective factor in agentic misuse settings, but it is fragile under minimal adversarial pressure, highlighting the need for personalization-aware evaluations and safeguards that remain robust across user-context conditions.
Abstract (translated)
URL
https://arxiv.org/abs/2603.16734
PDF
-
Infinite Problem Generator: Verifiably Scaling Physics Reasoning Data with Agentic Workflows
2026-03-15 17:08:50Aditya Sharan, Sriram Hebbale, Dhruv KumarAbstract
Training large language models for complex reasoning is bottlenecked by the scarcity of verifiable, high-quality data. In domains like physics, standard text augmentation often introduces hallucinations, while static benchmarks lack the reasoning traces required for fine-tuning. We introduce the Infinite Problem Generator (IPG), an agentic framework that synthesizes physics problems with guaranteed solvability through a Formula-as-Code paradigm. Unlike probabilistic text generation, IPG constructs solutions as executable Python programs, enforcing strict mathematical consistency. As a proof-of-concept, we release ClassicalMechanicsV1, a high-fidelity corpus of 1,335 classical mechanics problems expanded from 165 expert seeds. The corpus demonstrates high structural diversity, spanning 102 unique physical formulas with an average complexity of 3.05 formulas per problem. Furthermore, we identify a Complexity Blueprint, demonstrating a strong linear correlation ($R^2 \approx 0.95$) between formula count and verification code length. This relationship establishes code complexity as a precise, proxy-free metric for problem difficulty, enabling controllable curriculum generation. We release the full IPG pipeline, the ClassicalMechanicsV1 dataset, and our evaluation report to support reproducible research in reasoning-intensive domains.
Abstract (translated)
URL
https://arxiv.org/abs/2603.14486
PDF
-
Extending Minimal Pairs with Ordinal Surprisal Curves and Entropy Across Applied Domains
2026-03-15 14:31:00Andrew KatzAbstract
The minimal pairs paradigm of comparing model probabilities for contrasting completions has proven useful for evaluating linguistic knowledge in language models, yet its application has largely been confined to binary grammaticality judgments over syntactic phenomena. Additionally, standard prompting-based evaluation requires expensive text generation, may elicit post-hoc rationalizations rather than model judgments, and discards information about model uncertainty. We address both limitations by extending surprisal-based evaluation from binary grammaticality contrasts to ordinal-scaled classification and scoring tasks across multiple domains. Rather than asking models to generate answers, we measure the information-theoretic "surprise" (negative log probability) they assign to each position on rating scales (e.g., 1-5 or 1-9), yielding full surprisal curves that reveal both the model's preferred response and its uncertainty via entropy. We explore this framework across four domains: social-ecological-technological systems classification, causal statement identification (binary and scaled), figurative language detection, and deductive qualitative coding. Across these domains, surprisal curves produce interpretable classification signals with clear minima near expected ordinal scale positions, and entropy over the completion tended to distinguish genuinely ambiguous items from easier items.
Abstract (translated)
URL
https://arxiv.org/abs/2603.14400
PDF
-
Cheers: Decoupling Patch Details from Semantic Representations Enables Unified Multimodal Comprehension and Generation
2026-03-13 08:55:27Yichen Zhang, Da Peng, Zonghao Guo, Zijian Zhang, Xuesong Yang, Tong Sun, Shichu Sun, Yidan Zhang, Yanghao Li, Haiyan Zhao, Wang Xu, Qi Shi, Yangang Sun, Chi Chen, Shuo Wang, Yukun Yan, Xu Han, Qiang Ma, Wei Ke, Liang Wang, Zhiyuan Liu, Maosong SunAbstract
A recent cutting-edge topic in multimodal modeling is to unify visual comprehension and generation within a single model. However, the two tasks demand mismatched decoding regimes and visual representations, making it non-trivial to jointly optimize within a shared feature space. In this work, we present Cheers, a unified multimodal model that decouples patch-level details from semantic representations, thereby stabilizing semantics for multimodal understanding and improving fidelity for image generation via gated detail residuals. Cheers includes three key components: (i) a unified vision tokenizer that encodes and compresses image latent states into semantic tokens for efficient LLM conditioning, (ii) an LLM-based Transformer that unifies autoregressive decoding for text generation and diffusion decoding for image generation, and (iii) a cascaded flow matching head that decodes visual semantics first and then injects semantically gated detail residuals from the vision tokenizer to refine high-frequency content. Experiments on popular benchmarks demonstrate that Cheers matches or surpasses advanced UMMs in both visual understanding and generation. Cheers also achieves 4x token compression, enabling more efficient high-resolution image encoding and generation. Notably, Cheers outperforms the Tar-1.5B on the popular benchmarks GenEval and MMBench, while requiring only 20% of the training cost, indicating effective and efficient (i.e., 4x token compression) unified multimodal modeling. We will release all code and data for future research.
Abstract (translated)
URL
https://arxiv.org/abs/2603.12793
PDF
-
AI Model Modulation with Logits Redistribution
2026-03-13 07:57:23Zihan Wang, Zhongkui Ma, Xinguo Feng, Zhiyang Mei, Ethan Ma, Derui Wang, Minhui Xue, Guangdong BaiAbstract
Large-scale models are typically adapted to meet the diverse requirements of model owners and users. However, maintaining multiple specialized versions of the model is inefficient. In response, we propose AIM, a novel model modulation paradigm that enables a single model to exhibit diverse behaviors to meet the specific end requirements. AIM enables two key modulation modes: utility and focus modulations. The former provides model owners with dynamic control over output quality to deliver varying utility levels, and the latter offers users precise control to shift model's focused input features. AIM introduces a logits redistribution strategy that operates in a training data-agnostic and retraining-free manner. We establish a formal foundation to ensure AIM's regulation capability, based on the statistical properties of logits ordering via joint probability distributions. Our evaluation confirms AIM's practicality and versatility for Al model modulation, with tasks spanning image classification, semantic segmentation and text generation, and prevalent architectures including ResNet, SegFormer and Llama.
Abstract (translated)
URL
https://arxiv.org/abs/2603.12755
PDF
-
Beyond Sequential Distance: Inter-Modal Distance Invariant Position Encoding
2026-03-11 15:15:12Lin Chen, Bolin Ni, Qi Yang, Zili Wang, Kun Ding, Ying Wang, Houwen Peng, Shiming XiangAbstract
Despite the remarkable capabilities of Multimodal Large Language Models (MLLMs), they still suffer from visual fading in long-context scenarios. Specifically, the attention to visual tokens diminishes as the text sequence lengthens, leading to text generation detached from visual constraints. We attribute this degradation to the inherent inductive bias of Multimodal RoPE, which penalizes inter-modal attention as the distance between visual and text tokens increases. To address this, we propose inter-modal Distance Invariant Position Encoding (DIPE), a simple but effective mechanism that disentangles position encoding based on modality interactions. DIPE retains the natural relative positioning for intra-modal interactions to preserve local structure, while enforcing an anchored perceptual proximity for inter-modal interactions. This strategy effectively mitigates the inter-modal distance-based penalty, ensuring that visual signals remain perceptually consistent regardless of the context length. Experimental results demonstrate that by integrating DIPE with Multimodal RoPE, the model maintains stable visual grounding in long-context scenarios, significantly alleviating visual fading while preserving performance on standard short-context benchmarks. Code is available at this https URL.
Abstract (translated)
尽管多模态大型语言模型（MLLM）具有显著的能力，但在长上下文场景中仍然存在视觉衰减的问题。具体而言，随着文本序列长度的增加，对视觉标记的关注度逐渐减弱，导致生成的文字与视觉约束脱钩。我们认为这种退化是由于多模态RoPE的内在归纳偏差所导致的，在视觉和文本标记之间的距离增大时，会惩罚跨模态注意力。 为了解决这个问题，我们提出了跨模态距离不变位置编码（DIPE），这是一种简单但有效的机制，通过基于模态交互来分离位置编码。DIPE保留了单个模态内部互动中的自然相对定位以保持局部结构，并在跨模态互动中强制执行感知上的固定接近度。这一策略有效地缓解了基于跨模态距离的惩罚问题，确保视觉信号无论上下文长度如何都能保持感知上的一致性。 实验结果显示，通过将DIPE与多模态RoPE集成在一起，模型能够在长上下文中维持稳定的视觉定位，并显著减轻视觉衰减，同时在标准短上下文基准测试中保持性能。相关代码可在该链接获取：[此URL]（请使用实际的网址替代"this https URL"）。
URL
https://arxiv.org/abs/2603.10863
PDF
-
PET-F2I: A Comprehensive Benchmark and Parameter-Efficient Fine-Tuning of LLMs for PET/CT Report Impression Generation
2026-03-11 09:08:25Yuchen Liu, Wenbo Zhang, Liling Peng, Yichi Zhang, Yu Fu, Xin Guo, Chao Qu, Yuan Qi, Le XueAbstract
PET/CT imaging is pivotal in oncology and nuclear medicine, yet summarizing complex findings into precise diagnostic impressions is labor-intensive. While LLMs have shown promise in medical text generation, their capability in the highly specialized domain of PET/CT remains underexplored. We introduce PET-F2I-41K (PET Findings-to-Impression Benchmark), a large-scale benchmark for PET/CT impression generation using LLMs, constructed from over 41k real-world reports. Using PET-F2I-41K, we conduct a comprehensive evaluation of 27 models across proprietary frontier LLMs, open-source generalist models, and medical-domain LLMs, and we develop a domain-adapted 7B model (PET-F2I-7B) fine-tuned from Qwen2.5-7B-Instruct via LoRA. Beyond standard NLG metrics (e.g., BLEU-4, ROUGE-L, BERTScore), we propose three clinically grounded metrics - Entity Coverage Rate (ECR), Uncovered Entity Rate (UER), and Factual Consistency Rate (FCR) - to assess diagnostic completeness and factual reliability. Experiments reveal that neither frontier nor medical-domain LLMs perform adequately in zero-shot settings. In contrast, PET-F2I-7B achieves substantial gains (e.g., 0.708 BLEU-4) and a 3.0x improvement in entity coverage over the strongest baseline, while offering advantages in cost, latency, and privacy. Beyond this modeling contribution, PET-F2I-41K establishes a standardized evaluation framework to accelerate the development of reliable and clinically deployable reporting systems for PET/CT.
Abstract (translated)
PET/CT成像在肿瘤学和核医学领域至关重要，但将复杂的发现总结为精确的诊断结论是一项繁重的任务。尽管大语言模型（LLMs）在医疗文本生成方面显示出潜力，但在高度专业化的PET/CT领域中其能力尚待探索。我们推出了PET-F2I-41K (PET发现到印象基准)，这是一个使用大规模语言模型进行PET/CT印象生成的大规模基准测试数据集，构建自超过41,000份真实世界报告。利用PET-F2I-41K，我们在专有前沿大语言模型、开源通用模型和医学领域大语言模型中对27种不同的模型进行了全面评估，并开发了一个经过领域适应的7B模型（PET-F2I-7B），它是从Qwen2.5-7B-Instruct通过低秩适应（LoRA）微调而来。除了标准自然语言生成（NLG）指标（例如BLEU-4、ROUGE-L和BERTScore）之外，我们还提出了三种基于临床实践的度量标准——实体覆盖率率（ECR）、未覆盖实体率（UER）和事实一致性率（FCR），以评估诊断完整性和事实可靠性。实验结果显示，在零样本设置下，无论是前沿模型还是医学领域专用的大语言模型都无法充分表现。相比之下，PET-F2I-7B在诸如BLEU-4这样的标准指标上取得了显著的提升（如0.708 BLEU-4），并且在实体覆盖率方面比最强基线提高了3.0倍，同时提供了成本、延迟和隐私方面的优势。除了模型贡献外，PET-F2I-41K还建立了一个标准化评估框架，以加速可靠且临床可部署的报告系统的发展，用于PET/CT。
URL
https://arxiv.org/abs/2603.10560
PDF
-
LogoDiffuser: Training-Free Multilingual Logo Generation and Stylization via Letter-Aware Attention Control
2026-03-10 14:57:46Mingyu Kang, Hyein Seo, Yuna Jeong, Junhyeong Park, Yong Suk ChoiAbstract
Recent advances in text-to-image generation have been remarkable, but generating multilingual design logos that harmoniously integrate visual and textual elements remains a challenging task. Existing methods often distort character geometry when applying creative styles and struggle to support multilingual text generation without additional training. To address these challenges, we propose LogoDiffuser, a training-free method that synthesizes multilingual logo designs using the multimodal diffusion transformer. Instead of using textual prompts, we input the target characters as images, enabling robust character structure control regardless of language. We first analyze the joint attention mechanism to identify core tokens, which are tokens that strongly respond to textual structures. With this observation, our method integrates character structure and visual design by injecting the most informative attention maps. Furthermore, we perform layer-wise aggregation of attention maps to mitigate attention shifts across layers and obtain consistent core tokens. Extensive experiments and user studies demonstrate that our method achieves state-of-the-art performance in multilingual logo generation.
Abstract (translated)
近期在文本到图像生成领域取得了显著的进展，但在创建多语言设计标志时，如何和谐地整合视觉和文字元素仍然是一项挑战。现有的方法往往会在应用创意风格时扭曲字符几何形状，并且难以支持无需额外训练就能进行多种语言的文字生成。为了解决这些挑战，我们提出了一种名为LogoDiffuser的方法，这是一种无需训练的合成多语言标志设计的方法，使用的是多模态扩散变换器。不同于使用文本提示，我们输入目标文字作为图像，从而可以对任何语言的文字结构进行全面控制。 首先，我们通过分析联合注意力机制来识别核心令牌（即强烈响应于文本结构的令牌）。基于这一观察，我们的方法通过注入最具有信息量的注意力图谱将字符结构和视觉设计结合在一起。此外，我们执行层次化的注意力图谱聚合以减轻不同层之间的注意力偏移，并获取一致的核心令牌。 广泛的实验和用户研究证明了我们的方法在多语言标志生成方面达到了最先进的性能水平。
URL
https://arxiv.org/abs/2603.09759
PDF
-
VoxEmo: Benchmarking Speech Emotion Recognition with Speech LLMs
2026-03-09 21:10:34Hezhao Zhang, Huang-Cheng Chou, Shrikanth Narayanan, Thomas HainAbstract
Speech Large Language Models (LLMs) show great promise for speech emotion recognition (SER) via generative interfaces. However, shifting from closed-set classification to open text generation introduces zero-shot stochasticity, making evaluation highly sensitive to prompts. Additionally, conventional speech LLMs benchmarks overlook the inherent ambiguity of human emotion. Hence, we present VoxEmo, a comprehensive SER benchmark encompassing 35 emotion corpora across 15 languages for Speech LLMs. VoxEmo provides a standardized toolkit featuring varying prompt complexities, from direct classification to paralinguistic reasoning. To reflect real-world perception/application, we introduce a distribution-aware soft-label protocol and a prompt-ensemble strategy that emulates annotator disagreement. Experiments reveal that while zero-shot speech LLMs trail supervised baselines in hard-label accuracy, they uniquely align with human subjective distributions.
Abstract (translated)
语音大型语言模型（LLMs）在通过生成接口进行语音情感识别（SER）方面显示出巨大潜力。然而，从封闭集分类转向开放文本生成引入了零样本随机性，使得评估高度依赖于提示的敏感度增加。此外，传统的语音LLM基准测试忽略了人类情绪固有的模糊性。因此，我们提出了VoxEmo，这是一个全面的SER基准测试工具，涵盖了15种语言中的35个情感语料库，专为语音LLMs设计。VoxEmo提供了一个标准化的工具包，其中包括从直接分类到旁语义推理的各种提示复杂度。 为了反映真实世界的应用感知情况，我们引入了一种分布感知的软标签协议和一种提示集成策略，该策略模拟了标注人员之间的分歧意见。实验表明，虽然零样本语音LLMs在硬标签准确性上落后于监督基线模型，但它们独特地与人类主观分布相一致。
URL
https://arxiv.org/abs/2603.08936
PDF
