Power‑SMC is a training‑free sampling method that targets the sequence‑level power distribution used to sharpen large language model outputs for reasoning tasks. It maintains multiple candidate sequences, assigns importance weights based on how well each matches the power distribution, and periodically prunes low‑scoring candidates to keep inference latency close to standard decoding. Experiments on MATH500, GSM8K, GPQA, and HumanEval show that Power‑SMC matches or surpasses Metropolis–Hastings sampling in accuracy, preserves output diversity, and can accelerate inference speed by up to 17.6×.
By Seyedarmin Azizi, Erfan Baghaei Potraghloo, Minoo Ahmadi, Souvik Kundu, Massoud Pedram
The paper introduces <flowreader>, a method that casts evidence selection for long multimodal documents as a minimum‑cost flow problem over a multimodal content graph. It uses spectral decomposition to identify latent query‑relevant aspects and allocates a fixed evidence budget proportionally to their spectral energy, ensuring aspect coverage without a language‑model planning call. On the VisDoMBench benchmark with Qwen3‑VL‑32B, <flowreader> achieves the highest macro accuracy (68.9%) and outperforms prior systems by 2.7 points, while using fewer content nodes and reader tokens.
By Ambuj Mehrish, Sebastiano Vascon
The paper introduces XiangqiBench, an executable benchmark for evaluating large language model agents in Chinese chess. It tracks 8,568 multi‑turn trajectories from 12 frontier LLMs, revealing that metrics such as the Conversion Gap, Consistency Gap, and Simulation Gap overstate true closed‑loop success. The study shows that simply naming the correct move is insufficient; agents must reliably carry a plan through to a verified outcome against an opponent.
By Yekun Chai, Qiwei Peng, Haoyi Xiong
The paper investigates how post‑softmax attention weights behave across different language models by marginalizing over query positions to create a joint token‑head space. Across more than 60 diverse LLMs, token‑wise marginal attention reveals a text‑intrinsic signal that is conserved across models, while head‑wise marginal attention yields a model‑private signature that is consistent across documents. This insight enables a practical method for estimating per‑head budgets for key‑value cache eviction, achieving competitive performance on standard benchmarks without per‑document recomputation or training.
By Valentino Maiorca, Walter Nelson, Francesco Locatello
The paper introduces Collective Bias Mitigation (CBM), a framework that reduces bias in large language models by learning fine‑grained model behavior and enabling knowledge sharing among diverse LLMs. CBM systematically explores how to select and organize distinct models to produce fairer responses, outperforming standalone baselines. Experiments demonstrate significant bias reduction, with the Committee topology balancing effectiveness and inference cost.
By Mingzhe Du, Luu Anh Tuan, Xiaobao Wu, Yichong Huang, Yue Liu, Dong Huang, Huijun Liu, Bin Ji, Jie M. Zhang, See-Kiong Ng
StanceEval 2026 is the second edition of a shared task on stance detection in Arabic social media, where systems must classify a tweet’s stance toward a target as Favor, Against, or None. The event featured two tracks: Track 1 tests cross‑target transfer on thematically related topics (Women Driving vs. Women Empowerment), while Track 2 evaluates cross‑domain transfer to entirely unseen targets (E‑Cars and Trimester System). With 80 registered teams and 30 submissions, top systems achieved $F_{avg2}$ scores of 0.8994 (Track 1) and 0.9400 (Track 2), surpassing baseline performance and highlighting challenges such as target polarization and dialectal nuance.
By Rasha Albalawi, Nuha Albadi, Hamzah Luqman, Asma Yamani, Maram Kurdi, Saad Ezzini, Ahmed Ashraf, Maged Al-Shaibani, Nora Alturayeif
The paper investigates how large language model (LLM) agents exhibit source preference when selecting items in end‑to‑end search tasks across three domains. Across 12 agent models, each consistently favors certain sources and avoids others, even when items from disfavored sources better meet the user’s requirements. The study shows that revealing or hiding source identifiers influences selection, and that training rewards or missing information can create or mitigate these preferences.
By Jonghyun Song, Haewon Park, Jeonghoon Shim, Woojung Song, Yohan Jo
The paper introduces a single‑pass uncertainty head for detecting hallucinations in Persian medical language models. By adapting the LLM Uncertainty Head framework to Aya‑Expanse‑8B and two Gaokerena backbones, the authors build claim‑level hallucination datasets and train lightweight heads that achieve PR‑AUCs of 0.4820 and 0.4652 and ROC‑AUCs of 0.7852 and 0.7810. These heads operate without retrieval or repeated sampling during inference, offering a cost‑effective approach to hallucination detection in Persian medical LLMs.
By Mehrdad Ghassabi, Pedram Rostami, Hamidreza Baradaran Kashani, Sadra Hakim, Audrina Ebrahimi
The paper evaluates six structured decision models for hate‑speech moderation across four datasets, comparing them to specialized, zero‑shot, commercial, and supervised baselines. It investigates whether providing dataset definitions or decomposing them into multiple questions improves classification, and measures latency and cost. Results show that commercial LLMs outperform decision models on only one dataset, definitions alter up to 28% of predictions without consistent gains, and decomposition helps in only 20% of cases, yet the best hosted decision model approaches the best commercial LLM’s performance with roughly 97% lower inference cost.
By Demetris Paschalides, George Pallis, Marios D. Dikaiakos
The paper "Code2Math: Can Your Code Agent Evolve Math Problems Through Exploration?" explores how code agents can autonomously generate more complex variations of existing math problems. It introduces a multi‑agent framework that evolves problems while ensuring they remain solvable and increasingly difficult. Experiments show that with sufficient exploration, code agents can produce new, structurally distinct, and challenging problems, demonstrating a scalable method for creating high‑difficulty mathematical reasoning tasks.
By Dadi Guo, Yuejin Xie, Qingyu Liu, Weixian Huang, Jiayu Liu, Zhiyuan Fan, Qihan Ren, Shuai Shao, Tianyi Zhou, Jianjie Feng, Wenze Su, Yujiu Yang, Dongrui Liu, Yi R. Fung
The paper introduces Mawqif-XT, a new Arabic benchmark dataset comprising 996 manually annotated tweets from three public targets: Women Driving, E-Cars, and Trimester System. Each tweet is labeled for stance, sentiment, and sarcasm following the Mawqif annotation scheme, and the dataset is intended as a held‑out evaluation set to test cross‑target generalization. Baseline results are provided using Arabic and multilingual transformer models as well as zero‑shot large language models, enabling reproducible evaluation alongside the original Mawqif dataset.
By Rasha Albalawi, Nuha Albadi, Hamzah Luqman, Maram Kurdi, Saad Ezzini, Asma Yamani, Ahmed Ashraf
The paper introduces GAW-PO, a gradient‑aligned token reweighting technique for Direct Preference Optimization (DPO). It assigns weaker penalties to rejected tokens whose gradients align with preferred behavior, while maintaining stronger penalties for conflicting tokens. Across 11 benchmarks in mathematics, reasoning, coding, and question answering, GAW‑PO outperforms standard DPO and other baselines, and remains robust when the DPO regularization parameter is reduced.
By Andreea Dutulescu, Stefan Ruseti, Mihai Masala, Traian Rebedea, Mihai Dascalu
The paper introduces a simulation‑grounded vision‑language model (VLM) framework for wildfire monitoring that converts 2D wildfire simulations into labeled video episodes using a fixed Blender mapping to create low‑detail 3D proxies. These proxies, along with controllable video generation, provide a multimodal memory that a training‑free multi‑agent VLM system uses to retrieve reference episodes, reconcile visual and memory‑based predictions, and generate structured wildfire reports. The system achieves 77.3% accuracy on six simulator‑derived report fields, outperforming direct VLM querying and text‑only memory baselines.
By Duowen Chen, Yuchen Sun, Zhiqi Li, Yuxuan Liao, Sinan Wang, Bart van Bloemen Waanders, Bo Zhu
SpectralCache is a training‑free framework that accelerates diffusion‑based world models by exploiting the stability of singular subspaces and predictable evolution of singular values across denoising steps. It reuses these stable subspaces and estimates low‑dimensional singular values via linear extrapolation, while also skipping expensive backbone evaluations through singular value scaling. Experiments on models such as HunyuanWorld‑Voyager‑13B show a 5.22× speed‑up with no loss in generation quality, outperforming existing caching methods.
By Zhendong Mi, Pu Zhao, Ziyu Hu, Xiaodong Yu, Yanzhi Wang, Grace Li Zhang, Shaoyi Huang
The paper introduces “SpaceConflict”, a benchmark of 23,196 multimodal inputs designed to test whether large language models can not only report spatial facts but also use them in reasoning. Experiments show a gap between a model’s ability to recover an initial spatial state from visual evidence and its ability to apply that state to perform transformations, with the gap narrowing but not closing as model size increases. To address this, the authors propose Operational State Supervision (OSS), which supervises task‑relevant spatial states and their transformation trajectories, improving accuracy on judgments that require state organization and use.
By Jinchang Zhang, Guoyu Lu
The paper introduces Behavior Pack Optimization (BPO), a post‑training method for video multimodal large language models that replaces single‑response rewards with a set of outputs across counterfactual views. BPO enforces stability when interventions are irrelevant, sensitivity when key evidence is removed, and abstention when no evidence remains, using an anchor‑relative advantage to keep the objective stable with small pack sizes. Experiments on datasets such as TempCompass, MVBench, and NExT‑QA show that BPO improves macro accuracy and abstention metrics for models like Qwen2.5‑VL‑7B‑Instruct, with gains that transfer to other benchmarks and models.
By Zhaolu Kang, Shiyu Liu, Tailong Luo, Wei Zhang, Yingjie He, Lei Wei, Guansu Wang, Liang He, Siheng Wang, Guangyuan Dong, Jiaqi Su, Shuang Chen, Haoyu Ji, Qishi Zhan, Kaiyue Zhou
GeoScaffold introduces a geometric supervision framework that embeds 3D geometry into vision‑language navigation policies during training, eliminating the need for depth sensors or 3D encoders at inference. The method learns a compact depth tokenizer from training trajectories, then fine‑tunes the policy with geometry query tokens that reconstruct depth, connectivity, and traversability, turning these states into compact geometric latents for action decoding. After training, the tokenizer, generators, and reconstruction heads are discarded, leaving a lightweight backbone that outperforms vision‑only navigators on continuous VLN benchmarks.
By Yixuan Jiang, Wentong Li, An Liu, Zihao Xin, Fulin Tang, Cong Leng, Yang Gao, Jian Cheng
I2CD is a method that directly predicts a convex decomposition of 3D geometry from a single RGB image, bypassing the traditional slow reconstruct‑then‑decompose pipeline. It leverages a frozen pretrained diffusion transformer and shape decoder, adding only a lightweight cross‑attention head to output half‑plane parameters for K convex polytopes. The resulting geometry is compact, convex by construction, and can be loaded into physics engines in about 0.5 s per image, outperforming eight baseline pipelines in speed and volumetric IoU on benchmark datasets.
By Qian Wang, Liam Merz Hoffmeister, Brian Scassellati, Daniel Rakita
The paper investigates how affective fine‑tuning shapes the internal architecture of multimodal foundation models. By analyzing 13 model instances across nine designs, it finds that adapting the feed‑forward network (FFN) consistently outperforms attention‑only adaptation and nearly matches full‑model tuning, revealing the FFN as an efficient adaptation substrate. Moreover, joint optimization leads to emergent functional specialization, notably a prominent gate projection pathway, which the authors exploit in Gate‑Focused Efficient Tuning (GET) to achieve 96.2–98.0% of full‑model performance with only 19.3–24.5% of the trainable parameters.
By Zhen Zhang, Runhao Zeng, Sicheng Zhao, Xiping Hu
LiteTrajEval is a lightweight architecture designed for budget‑bounded trajectory evaluation of large‑language‑model agents. It creates compact domain‑specific rule profiles offline, preprocesses trajectories online to flag heuristic failure signals, and then uses a single rubric‑guided LLM judge to generate structured diagnostic reports. On public Magnetic‑One‑style and τ‑bench‑style datasets, LiteTrajEval improves failure‑localization alignment with human annotations by 20–35 percentage points and reduces cost and evaluation time by 6× and 8× respectively, and has been deployed in an enterprise agentic platform.
By Linh-An Phan, MingXue Wang, Guangyu Wu, Feng Pan, Zhaoyu Pang, Yanbin Zhang