论文检索

Pengyuan Qin, Linnan Tu, Yuhan Ke, Hefei Ling

Large language models (LLMs) watermarking has been proposed as an active approach for content provenance verification, yet existing evaluations are largely confined to fixed entropy settings. In this paper, we introduce EntroBench, a benchmark for LLM watermarking that systematically covers three entropy levels and seven representative tasks. We conducted a fair evaluation of eight watermarking methods through hyper-parameter search based on an anchored dataset. We find that current approaches struggle to perform consistently across different entropy levels. Our analysis reveals a clear trade-off between watermark detectability and downstream output quality that varies across tasks and entropy conditions. Furthermore, we assess watermark robustness under realistic user interaction scenarios and show that common, non-adversarial user behaviors can substantially degrade watermark signals. These results indicate that practical usage-driven perturbations pose a significant challenge to current watermarking techniques. EntroBench provides a unified evaluation framework for studying these issues and supports the development of more adaptive and robust LLM watermarking methods. Dataset and codes are available at https://github.com/py-qin/EntroBench.

Hong Kyu Lee, Ruixuan Liu, Li Xiong

Machine unlearning is an emerging technique that removes the influence of a subset of training data (forget set) from a model without full retraining, with applications including privacy protection, content moderation, and model correction. The key challenge lies in achieving strong unlearning efficacy while preserving the overall utility. Existing unlearning methods for large language models (LLMs) often rely on auxiliary models, retain datasets, or even commercial AI services. However, dependence on these external resources is often impractical and could potentially introduce additional privacy risks. In this work, we propose direct token optimization (DTO), a self-contained unlearning approach for LLMs that directly optimizes the token-level objectives to unlearn specific sequences without external resources.For each sequence to be unlearned, we identify target tokens that encode critical knowledge for unlearning and treat remaining tokens as non-target ones for maintaining the model utility. DTO maximizes an unlearning objective on target tokens and applies a utility-preservation regularizer on non-target tokens.Across multiple unlearning benchmarks, DTO improves the forget quality up to 16.8\times over the latest baselines while maintaining comparable model utility. Our code is available at github.com/Emory-AIMS/direct_token_optimization.

Aakriti Agrawal, Gouthaman KV, Rohith Aralikatti, Gauri Jagatap, Jiaxin Yuan, Sarvesh Baskar, Vijay Kamarshi, Andrea Fanelli, Furong Huang

Hallucinations in Large Vision-Language Models (LVLMs) remain a persistent challenge, often stemming from inadequate integration of visual information during multimodal reasoning. A key cause is the model’s over-reliance on textual priors and underutilization of visual cues, leading to outputs that are linguistically fluent but visually inaccurate. For example, given an image of an empty kitchen countertop, an LVLM might hallucinate a “bowl of fruit” or “cup of coffee,” relying on language associations rather than visual evidence. Most LVLMs incorporate visual features by appending them to the input stream of a pre-trained LLM and training on large-scale vision-language datasets. Our systematic analysis reveals that this strategy often leads to over-dependence on textual information due to the inherent bias of LLMs towards language-dominant representations. This imbalance skews attention towards the text over visual content, weakening the model’s ability to ground outputs in visual inputs. To address this, we propose a simple yet effective visual feature incorporation method that encourages the model to learn visually-informed textual embeddings distinct from those of the base LLM and promotes a more balanced attention distribution. Experimental results across multiple hallucination benchmarks demonstrate that our method significantly reduces hallucinations and fosters more balanced multimodal reasoning. Notably, our approach achieves substantial gains, including \textbf{+9.33}% on MMVP-MLLM, \textbf{+2.99}% on POPE-AOKVQA, up to \textbf{+3.4}% on Merlin, and \textbf{+3}% on the hard-data split of HallusionBench.

Neeva Oza, Ishaan Govil, Parul Gupta, Dinesh Khandelwal, Dinesh Garg, Parag Singla

We study how well LLMs can determine whether two programs are functionally equivalent. This is an important problem because benchmarking code equivalence helps assess LLM capability in tasks such as code rewriting and translation. To this end, we introduce CETBench — Code Equivalence with Transformations Benchmark — built from a repository of programs that may solve the same or different tasks. Each dataset instance is created by sampling a program pair and applying a random sequence of predefined code transformations, yielding either equivalent or non-equivalent pairs. Our analysis shows that even simple transformations cause a significant performance drop in state-of-the-art LLMs on code-equivalence checking. These challenges are further amplified in the cross-lingual setting when comparing programs written in different languages. To remedy this, we present a simple fine-tuning-based approach to boost LLM performance on the transformed pairs of programs. Our approach for dataset generation is generic, supporting cross-lingual equivalence checking, the generation of program pairs with varying difficulty levels, and the application of diverse transformations. In our experiments, we perform ablations over the difficulty level of original programs, as well as the kind of transformations used in generating pairs for equivalence checking. Our analysis presents deep insights into the working of LLMs for the task of code-equivalence, and points to the fact that they may still be far from what could be termed as a semantic understanding of the underlying code.

Hanna Yukhymenko, Anton Alexandrov, Martin Vechev

The reliability of multilingual Large Language Model (LLM) evaluation is currently compromised by the inconsistent quality of translated benchmarks. Existing resources often suffer from semantic drift and context loss, which can lead to misleading performance metrics. In this work, we present a fully automated framework designed to address these challenges by enabling scalable, high-quality translation of datasets and benchmarks. We demonstrate that adapting test-time compute scaling strategies, specifically Universal Self-Improvement (USI) and our proposed multi-round ranking method, T-RANK, allows for significantly higher quality outputs compared to traditional pipelines. Our framework ensures that benchmarks preserve their original task structure and linguistic nuances during localization. We apply this approach to translate popular benchmarks and datasets into eight Eastern and Southern European languages (Ukrainian, Bulgarian, Slovak, Romanian, Lithuanian, Estonian, Turkish, Greek). Evaluations using both reference-based metrics and LLM-as-a-judge show that our translations surpass existing resources, resulting in more accurate downstream model assessment. We release both the framework and the improved benchmarks to facilitate robust and reproducible multilingual AI development.

Rishabh Kumar, Abhinav Painuli, Chriss Philip Saji, Devesh Soni, Amrith Krishna, Ganesh Ramakrishnan

Quality management when creating large-scale speech datasets is essential for building reliable downstream models, yet verification pipelines are often brittle, domain-specific, and expertise-intensive. We introduce SpeechQM-Agent, a natural language-driven agentic framework that compiles user requirements into dependency-aware DAG workflows over modular tools for audio, transcript, and metadata verification. A central planner LLM enforces prerequisites and supports execution-time replanning (e.g., re-running failed steps or swapping tools), reducing manual pipeline engineering and improving robustness across heterogeneous vendor formats and multilingual settings. We also release SpeechQM-Dataset, a multilingual benchmark with controlled, vendor-inspired quality artifacts spanning 24 verification tasks. Across experiments, SpeechQM-Agent attains 80-90% agreement with expert verification while requiring <20% of the cost and time of manual QC, and we further validate transfer to real vendor-supplied corpora. Planner LLM comparisons highlight fidelity-efficiency trade-offs.

Xingyou Yin, Ceyao Zhang, Min Hu, Kai Chen

Large Language Models (LLMs) have demonstrated effectiveness as zero-shot time series (TS) forecasters. While existing work often relies on fine-tuning specialized modules to bridge this gap, a distinct, yet challenging, paradigm aims to leverage truly off-the-shelf LLMs without any fine-tuning whatsoever, relying solely on strategic tokenization of numerical sequences. However, the parameters of these fully frozen models cannot adapt to distribution shifts. Thus, we introduce a novel yet highly effective strategy to overcome this brittleness: injecting noise into the raw TS before tokenization. This non-invasive intervention acts as a form of inference-time augmentation, compelling the frozen LLM to extrapolate based on robust underlying temporal patterns rather than superficial numerical artifacts. We theoretically analyze this phenomenon and empirically validate its effectiveness across diverse benchmarks. Notably, to fully eliminate potential biases from data contamination during LLM pre-training, we introduce multiple novel real-world TS datasets that fall outside all utilized LLMs’ pre-training scopes, and consistently observe improved performance. This study provides a further step in directly leveraging off-the-shelf LLMs for TS forecasting[<https://github.com/jkumh/NLTS>].

Bo Zheng, Yudong Chen, Zihua Xiong, Shuai Fang, Peidong He, Yang Yang, Sheng Guo

Tabular data forms the backbone of high-stakes decision systems in finance, healthcare, and beyond. Yet industrial tabular datasets are inherently difficult: high-dimensional, riddled with missing entries, and rarely labeled at scale. While foundation models have revolutionized vision and language, tabular learning still leans on handcrafted features and lacks a general self-supervised framework. We present MaskTab, a unified pre-training framework designed specifically for industrial-scale tabular data. MaskTab encodes missing values via dedicated learnable tokens, enabling the model to distinguish structural absence from random dropout. It jointly optimizes a hybrid supervised pre-training scheme—utilizing a twin-path architecture to reconcile masked reconstruction with task-specific supervision—and an MoE-augmented loss that adaptively routes features through specialized subnetworks. On industrial-scale benchmarks, it achieves +5.04% AUC and +8.28% KS over prior art under rigorous scaling. Moreover, its representations distill effectively into lightweight models, yielding +2.55% AUC and +4.85% KS under strict latency and interpretability constraints, while improving robustness to distribution shifts. Our work demonstrates that tabular data admits a foundation-model treatment—when its structural idiosyncrasies are respected.

Saket Reddy, Ke Yang, ChengXiang Zhai

Mitigating social bias in Large Language Models (LLMs) presents a distinct alignment challenge: unlike verifiable tasks, social bias lacks a single ground truth, creating a high-variance, subjective reward landscape. Previous preference-based fine-tuning methods have major trade-offs: Direct Preference Optimization (DPO) is limited by the lack of exploration inherent in offline training, while Proximal Policy Optimization (PPO) can lead to training instability due to potentially unreliable critic estimates. In this paper, we propose BiasGRPO, an adaptation of Group Relative Policy Optimization (GRPO) that stabilizes alignment by normalizing rewards across a group of sampled completions. By substituting the value function with a group-relative baseline, our approach reduces instability while maintaining the exploration benefits of online reinforcement learning. To adapt GRPO, we curate and synthetically extend a dataset spanning multiple domains and contexts, and create a custom, bias-specific reward model for effectively guiding generation while avoiding knowledge degradation. We find that BiasGRPO outperforms DPO and PPO across multiple benchmarks, indicating its effectiveness as an alignment technique that can overcome the limitations of previous preference-based methods.

Dingyi Yang, Mingshuo Wang, Qin Jin

Automatic evaluation of book-length stories remains underexplored, particularly for non-English literature. We introduce ChangJuan, the first benchmark for *book-length Chinese story evaluation*, comprising 300 novels with metadata, human ratings, and large-scale user reviews. To mitigate the subjectivity of raw reviews, we propose a distillation method to aggregate them into generally agreed viewpoints (pros and cons) across key evaluation aspects such as plot and character. We conduct systematic experiments to benchmark current LLMs, analyze aspect importance, and examine genre differences. For book-length story evaluation, we propose an enhanced summary-based method that leverages length-detail balanced summaries and representative excerpts, generates aspect-specific reviews, and considers genre-aware aspect weighting to assign a final score. Using this framework and our distilled viewpoints, we fine-tune an 8B model, CLEM, which outperforms open-source baselines and raises Qwen3’s Kendall’s tau correlation with human judgments from 24.8 to 34.1. Our datasets and codes are available at https://github.com/DingyiYang/ChangJuan.

Chenye Zhao, Cornelia Caragea

Stance detection aims to ascertain whether an author’s text is in favor, against, or neutral toward specific targets like public policies or social issues. While pretrained language models (PLMs) have greatly enhanced stance detection, they remain vulnerable to adversarial attacks—manipulations that maintain textual semantics but lead to incorrect predictions. Such vulnerabilities remain underexplored for stance detection. In this study, we introduce StanceAttack, an innovative adversarial attack method leveraging ChatGPT to create adversarial examples that can mislead well-trained stance detection models. We conduct experiments to evaluate our attack model by attacking state-of-the-art PLMs on two benchmark datasets. Results demonstrate that StanceAttack outperforms traditional adversarial methods with higher success rates and fewer retries. Human evaluations confirm that our adversarial examples preserve the original semantic meanings and naturalness. We share our code and data in https://github.com/chenyez/StanceAttack.

Jiazhou Liang, Yifan Simon Liu, David Guo, Yilun Jiang, Minqi Sun, Scott Sanner

The growing ubiquity of Extended Reality (XR) is driving Conversational Recommendation Systems (CRS) toward visually immersive experiences. We formalize this paradigm as Immersive CRS (ICRS), where recommended items are highlighted directly in the user’s scene-based visual environment and augmented with in-situ labels. While item recommendation has been widely studied, the problem of how to select and evaluate which information to present as immersive labels remains an open problem. To this end, we introduce a principled categorization of information needs into explicit intent satisfaction and proactive information needs and use these to define novel evaluation metrics for item label selection. We benchmark IR-, LLM-, and VLM-based methods across three datasets and ICRS scenarios: fashion, movie recommendation, and retail shopping. Our evaluation reveals three important limitations of existing methods: (1) they fail to leverage scenario-specific information modalities (e.g., visual cues for fashion, metadata for retail), (2) they present redundant information that is visually inferable, and (3) they poorly anticipate users’ proactive information needs from explicit dialogue alone. In summary, this work provides both a novel evaluation paradigm for in-situ item labeling in ICRS and highlights key challenges for future work.

Xinyi Zhao, Haoqi Hu, Ziyu Wang, Jinfeng Xiao

Large language models (LLMs) are increasingly used as judges to evaluate, rank, and supervise other models, yet their reliability in judging LLMs’ reasoning process under long-context settings remains underexplored. Existing benchmarks either overly rely on human annotators, who may miss subtle flaws in lengthy reasoning chains, or focus solely on final responses while ignoring the underlying context and reasoning process. We introduce Long-Reason Bench (LRBench), a large-scale benchmark for evaluating LLM-based judges. LRBench comprises over 100K annotated instances spanning medical, legal, and academic-review scenarios, with fine-grained labels indicating violations of six core principles: Logical Correctness, Factual Consistency, Bias and Fairness, Groundedness, Helpfulness, and Harmlessness. Experimental results reveal that state-of-the-art LLM judges struggle to identify nuanced reasoning errors in long contexts. To improve judge reliability, we further present Judge-R1, which combines reinforcement learning with multi-turn search to enable grounded and principle-aware evaluation. Across domains and principles, Judge-R1 consistently outperforms single-turn baselines, enabling scalable and trustworthy evaluation of LLM reasoning. Our dataset and code are available at https://github.com/Xinyi-0724/Judge-R1.

Shipeng Li, Zhiqin Yang, Shikun Li, Xiaobo Xia, Hengyu Liu, Xinghua Zhang, Gaode Chen, Dong Fang, Ying Tai, Zhe Peng

Reinforcement learning with verifiable rewards (RLVR) has become a key technique for enhancing LLMs’ reasoning abilities, yet its data inefficiency remains a major bottleneck. To address this critical yet challenging issue, we present a novel gradient-alignment-based method, named LearnAlign, which intelligently selects the learnable and representative training reasoning data for RLVR post-training. To overcome the well-known response-length bias in gradient norms, we introduce the data learnability based on the success rate, which indicates the learning potential of each data point. Experiments across five reasoning benchmarks show that our method significantly reduces training data requirements while achieving minor performance degradation or even improving performance compared to full-data training. Specifically, it reduces data requirements by up to 1,000 data points with better performance (77.5%) than that on the full dataset on the GSM8K benchmark (77.0%). Furthermore, its efficiency is demonstrated on both mathematical and code benchmarks by using much less data from the DAPO-MATH-17K dataset.

Saransh Sharma, Pritika Ramu, Aparna Garimella, Koyel Mukherjee

Answering open-ended questions remains challenging for AI systems because it requires synthesis, judgment, and exploration beyond factual retrieval, and users often refine answers through multiple iterations rather than accepting a single response. Existing QA benchmarks do not explicitly support this refinement process. To address this gap, we introduce a new task, document-grounded related insight generation, where the goal is to generate additional insights from a document collection that help improve, extend, or rethink an initial answer to an open-ended question, ultimately supporting richer user interaction and a better overall question answering experience. We curate and release SCOpE-QA (Scientific Collections for Open-Ended QA), a dataset of 3,000 open-ended questions across 20 research collections. We present InsightGen, a two-stage approach that first constructs a thematic representation of the document collection using clustering, and then selects related context based on neighborhood selection from the thematic graph to generate diverse and relevant insights using LLMs. Extensive evaluation on 3,000 questions using two generation models and two evaluation settings shows that InsightGen consistently produces useful, relevant, and actionable insights, establishing a strong baseline for this new task.

Aida Sanatizadeh, Sorouralsadat Fatemi, Reza Mousavi, Ahmed Abbasi

Large Language Models (LLMs) are increasingly used for structured tabular data, yet it remains unclear whether their performance reflects genuine reasoning or memorization of pre-training corpora. We investigate this question through a rigorous, contamination-aware evaluation of a representative modular Multi-Agent LLM (MALLM) framework against state-of-the-art AutoML systems and established baselines (TABLET, TABLLM). We evaluate eleven binary classification tasks: five pre-cutoff benchmarks likely seen during LLM pre-training and six post-cutoff datasets released after the LLM knowledge cutoff. Results show a sharp performance dichotomy: MALLM achieves competitive or superior performance on pre-cutoff datasets but substantially underperforms AutoML on post-cutoff data, exhibiting poor calibration and high variance, especially on hard-to-classify instances. By contrast, AutoML models generalize consistently and align confidence more closely with instance hardness. These findings suggest that, despite agentic scaffolding, current LLMs cannot yet replace production-grade discriminative models for tabular classification, underscoring the need for contamination-free benchmarks to accurately assess tabular reasoning capabilities.

Menglan Chen, Xianghe Pang, Jingjing Dong, WenHao Wang, Yaxin Du, Siheng Chen

Aligning Vision-Language Models (VLMs) with safety standards is essential to mitigate risks arising from their multimodal complexity, where integrating vision and language unveils subtle threats beyond the reach of conventional safeguards. Inspired by the insight that reasoning across modalities is key to preempting intricate vulnerabilities, we propose a novel direction for VLM safety: multimodal reasoning-driven prompt rewriting. To this end, we introduce VLMGuard-R1, a proactive framework that refines user inputs through a reasoning-guided rewriter, dynamically interpreting text-image interactions to deliver refined prompts that bolster safety across diverse VLM architectures without altering their core parameters. To achieve this, we devise a three-stage reasoning pipeline to synthesize a dataset that trains the rewriter to infer subtle threats, enabling tailored, actionable responses over generic refusals. Extensive experiments across five benchmarks with six VLMs reveal that VLMGuard-R1 outperforms four baselines. In particular, VLMGuard-R1 achieves a remarkable 43.59% increase in average safety across five models on the SIUO benchmark.

Xuanao Huang, Xingjia Liu, Zetong Zhou, Yuyang Peng, Yao Wan, Dongping Chen

Knowledge about the visual world is not only constantly evolving but also inherently happening all over the world: breaking news in Tokyo, political events in São Paulo, and cultural phenomena in Cairo are first reported in Japanese, Portuguese, and Arabic, carrying regional context that English-centric resources cannot fully capture. Yet existing resources for visual knowledge remain confined to English, creating a "Worldwide Knowledge Gap" that hinders developing truly global assistants. To quantify this gap, we introduce LiveVQA-W(orldwide), the first dynamic-updating dataset for real-time, multilingual visual knowledge seeking and updating across ten major languages. Drawing from worldwide news outlets, YouTube videos, and academic platforms during August–December 2025, LiveVQA-W comprises 234K images, 873K questions, and 171K visual entities with hierarchical evaluation: Level 1 for visual entity recognition and Level 2 for multi-hop cross-lingual reasoning. Our comprehensive benchmarking of 15 state-of-the-art MLLMs reveals that models without search achieve near-random performance, while search-augmented models exhibit severe linguistic bias, with English accuracy nearly double that of other languages. Furthermore, we explore visual knowledge updating through large-scale training, finding that injected knowledge improves recall but remains fragile under prompt rephrasing and image perturbations such as rotation and flipping. We release the fully replicable data collection pipeline and raw dataset to support continuous community-driven expansion. The benchmark, code, and related resources are available at: https://worldwide-livevqa.github.io.

Haebin Shin, Lei Ji, Xiao Liu, Zhiwei Yu, Hyunwoo Yoo, Qi Chen, Yeyun Gong

As numerous instruction-tuning datasets continue to emerge, dynamically balancing and optimizing their mixtures has become a criticalchallenge. To address this, we propose DynamixSFT, a dynamic and automated method for instruction-tuning dataset mixture optimization. We formulate the problem as a multi-armed bandit setup and introduce a Prior-scaled Boltzmann Exploration that softly anchors the updated sampling distribution to the original dataset proportions, thereby preserving the inherent diversity and coverage of the collection. Sampling probabilities are updated using a lightweight 1-Step Look-ahead Reward, reflecting how much the dataset contributes to improving the model’s performance at its current state. We demonstrate that DynamixSFT effectively optimizes the TÜLU-2-mixture andTÜLU-3-mixture collections across 10 benchmarks, while introducing minimal computational overhead over naive sampling. Furthermore, we provide a comprehensive analysis and visualizations to offer deeper insights into the adaptive dynamics of our method.

Mengyi Deng, Zhiwei Li, Xin Li, Tingyu Zhu, Yulan Yuan, Zhijiang Guo, Wei Wang

Although Large Language Models (LLMs) have made remarkable progress, current preference optimization methods still struggle to align directional consistency while preserving reasoning diversity. To address this limitation, we propose Directional-Groupwise Preference Optimization (DGPO), a lightweight framework that aggregates supervision signals at the group level and explicitly models direction-aware alignment through multi-candidate comparisons. DGPO organizes forward and reverse question-answer instances into structured sets and optimizes a margin-based likelihood objective that separates coherent reasoning paths from inconsistent alternatives. This groupwise formulation captures richer relative information than pairwise objectives and reinforces consistency across diverse reasoning pathways. Empirical results show that our constructed reverse data yields a 3.2% average improvement across five benchmarks, while DGPO further delivers consistent gains across multiple datasets and model families, achieving average accuracy improvements of up to 3.6%. Our code and data are available at https://github.com/Demi-deng2/DGPO.