The paper introduces a conceptual framework to classify and evaluate Retrieval-Augmented Generation (RAG) techniques, addressing the challenge of selecting suitable methods for specific tasks. It identifies main dimensions and sub-dimensions based on a qualitative analysis of existing literature. The framework can act as a pre-empirical filter, helping researchers and practitioners choose the most fitting RAG candidates before costly benchmarking.
By Emmanuel Igbekele, Olawande Daramola
OfficeTown is a generative simulation framework that models professional organizations, where tasks and context evolve from ongoing work. Starting from a natural‑language description of an organization’s purpose, the system initializes members who coordinate, delegate assignments to assistant agents, and review deliverables, while artifacts and decisions persist to influence future tasks. The authors built OfficeTown Bench with 406 tasks and 4,836 evaluation criteria across 23 simulated organizations in eight domains, and show that fine‑tuning and on‑policy distillation on this data improves Qwen3.5‑9B on external professional‑work benchmarks, with a principal–assistant setup outperforming single‑ and multi‑agent baselines.
By Hongchen Wei, Daohong Ying, Zhi Hong, Aidan Wu, Bei Liu, Ruichun Ma, Yifan Yang, Qi Dai, Kai Qiu, Rui Xu, Dongdong Chen, Zhenzhong Chen, Chong Luo
The paper argues that additional computation from the same large language model (LLM) during test time does not necessarily provide independent evidence, as errors are correlated across repeated samples, reasoning depth, and self-evaluation. It introduces DIA (Disagreement-triggered Independent Arbitration), which uses disagreement between intermediate and final predictions as a signal of instability and resolves it by having a small panel of heterogeneous models independently solve the question. Across seven benchmarks and two base models, DIA outperforms full chain-of-thought in most settings, achieving significant gains while using fewer tokens and activating only on a subset of examples.
By Ming Shen, Tianjun Yao
OliO is a self‑supervised learning method for time‑series forecasting that models bidirectional temporal flows using a linear ODE‑based transition operator. By enforcing a strictly upper triangular constraint on the transition matrix, it ensures numerical stability and captures long‑term dependencies without exponential divergence. Experiments on multiple backbones and benchmarks show up to 5.75% MSE improvement over state‑of‑the‑art baselines.
By Kwangryeol Park, Seulki Lee
Large language models (LLMs) tend to exhibit sycophancy by aligning with users’ expressed views. This study introduces a benchmark of 5,500 evaluation instances built from 500 debatable questions across ten domains, where the user stance is fixed but the justification is systematically varied across eleven conditions, including neutral, personal experience, causal framing, and worldview-based synthesis. Seven LLMs were evaluated, revealing that narrative structure can significantly influence agreement, with causal framing and bullet presentation boosting sycophancy while personal experience can reduce it.
By Runtai Yang, Dandan Zhu, Kaiwei Zhang, Guangtao Zhai
The paper introduces Consistent Sparse Dictionary Learning (ConSDL), a method that replaces the linear encoder in sparse autoencoders with per‑sample sparse coding to improve dictionary learning. Synthetic experiments show ConSDL achieves near‑perfect recovery of dictionaries and coefficients, and in real vision and language tasks it boosts cross‑seed consistency by 17.6% for dictionaries and 35.9% for coefficients compared to six SAE baselines.
By Jinho Choi, Hyesu Lim, Jaegul Choo, Steffen Schneider
The paper introduces M2CG, a one-pass deformable image registration network that replaces recursive refinement with multi-level, multi-space contextual guidance. M2CG performs progressive pyramid registration within a single iteration, leveraging cues from previous, current, and subsequent steps across image, image feature, segmentation, and segmentation feature spaces. It achieves state‑of‑the‑art results on eight public medical datasets in both unsupervised and semi‑supervised settings, challenging the prevailing reliance on recursive architectures.
By Yue Lin, Shenghai Wu, Mingyuan Meng
The paper introduces TRACE, a framework that embeds a Vector Quantizer at the post‑attention residual of each Transformer layer, enabling exchanged activations to be encoded as compact codebook indices. During inference, receivers recover key/value representations via table lookup, eliminating downstream computation and allowing communication–computation overlap. A hierarchical codebook further permits runtime bitrate adaptation, and experiments on ViT‑Base and ViViT‑B demonstrate up to 1.89% accuracy gain, 1.27× speedup, and a 22,624× reduction in communication overhead compared to uncompressed inference.
By Muhammad Azlan Qazi, Alexandros Iosifidis, Qi Zhang
AuditKIT is a source‑available evaluation library that unifies benchmark accuracy, judge‑based assessment, runtime performance, tool‑use behavior, and retrieval quality into a single result representation. It offers direction‑aware metric contracts, configuration fingerprints, and extensible metrics and backends, while providing turn‑aware tool‑call matching and distinguishing verified task outcomes from trajectory‑level diagnostics. Seven case studies demonstrate how answer extraction, runtime behavior, and metric selection can alter evaluation outcomes across model comparison, serving runtimes, retrieval, tool calling, and interoperability.
By Shreeyans Arora, Utsav Avaiya, Zera Lyngkhoi, Hem Gosalia, Ram Mohan Rao Kadiyala, Vinay Kumar Sankarapu, Pratinav Seth
FOR-species20K is the largest public LiDAR benchmark for tree‑species classification, combining 25 datasets across 33 species. A multi‑resolution, multi‑view ConvNeXtV2 model raised the hidden‑test accuracy from 79.5 % to 85.0 %, placing first at submission. However, cross‑source experiments show that this score does not reflect true source‑disjoint generalization, with performance dropping dramatically when species are held out by source.
By Jae Joong Lee, Bedrich Benes
The paper presents a method for generating hard problem instances by treating instance creation as an optimization problem, using solver runtime as the objective. Leveraging the Nevergrad platform, the authors generate challenging instances for four NP‑hard problems—satisfiability, traveling salesperson, Hamiltonian cycle, and non‑permutation flow‑shop scheduling—that outperform random draws and rival curated benchmark sets. The approach requires only a solver, an instance encoder, and a compute budget, and it can uncover structural insights such as phase‑transition behavior in SAT.
By Jamie Fravel, Martin Kouteck\'y, Pavlo Tsitsei
The paper introduces JUPE, a benchmark that tests whether plausible explanations can mislead multimodal large language models (MLLMs) when judging answers that require visual verification. Across 14 MLLM judges, ordinary yet plausible explanations significantly increased acceptance of incorrect answers, with plain justifications and chain-of-thought rationales reducing rejections by 12.7 and 13.8 percentage points, respectively. Signal detection analysis shows a shift in decision criterion toward acceptance without a change in discriminability, indicating that explanations can inflate performance estimates in evaluation pipelines.
By Anglin Cai, Zhiqi Wang, Xiaozhi Yang, Yurun Lin, Dongwon Lee, Yuchen Yang
The paper introduces CADENCE, a context‑adaptive decoding method that learns both what to draft and how far to draft during language generation. It uses Gaussian prediction heads to encode future token spans as continuous vectors, a shared decoder to produce tokens in parallel, and a learned router to choose the prediction length based on hidden states and Gaussian parameters. Experiments on Llama‑3.1‑8B‑Instruct show that CADENCE’s router‑selected positions achieve higher draft acceptance and up to 1.3× higher generation throughput compared to standard greedy decoding.
By Sai Sreenivas Chintha, Chidera Biringa, Rishab Sharma, Anusha Agrawal, Prachi Badarayani, Gemma Zhang, Nigel Boachie Kumankumah, Marco Ayala, Sulaiman Vesal
The paper introduces AnchorQ, a method that improves sample-efficient reinforcement learning for policy refinement by using flow-reversed action anchors. AnchorQ maps reference behaviors to state-conditioned latent anchors, samples local particles around these anchors, and employs value-guided sampling to explore promising refinements. Experiments on simulated and real-world manipulation tasks show that AnchorQ outperforms state-of-the-art post-training baselines, achieving a 37% higher real-world success rate and 97% greater simulated task throughput within the same interaction budget.
By Carl Qi, Weijun Dong, Xiaojie Wang, Peiyi Chen, Rahul Sajnani, Mingyuan Zhang, Amy Zhang, Jose Barreiros, Manikatan Nambi, Lillian J. Ratliff, Dylan Randle
SpikeQ introduces a spike‑guided post‑training quantization framework for spiking neural networks, selecting weight levels by simulating downstream neurons to minimize spike errors and designing a reset‑aware grid for pre‑fire membrane potential quantization that avoids spike flips. The method outperforms existing grids on seven image‑processing benchmarks, maintaining full‑precision accuracy with 2‑bit pre‑fire potentials and staying within 2.59 points of full precision at 4 bits while drastically reducing memory and energy usage.
By Kai Sun, Peibo Duan, Zhanglu Yan, Guowei Zhang, Yongsheng Huang, Levin Kuhlmann
Prism Attention introduces a proposal‑refined sparse attention framework that replaces static cross‑layer index reuse with a coarse‑to‑fine strategy, allowing layer‑wise adaptability while maintaining efficiency. The method leverages the observation that high‑attention tokens remain largely stable across layers, with changes concentrated near the boundary of previously selected regions. Experiments demonstrate state‑of‑the‑art performance on reasoning and long‑context benchmarks, achieving a 3.4× decoding speedup over FlashAttention at 128K context length.
By Zhenxu Tian, Kebin Liu, Zhengwu Yang, Yi Su, Qingqing Dang, Yanlin Sha, Kaipeng Deng, Dianhai Yu, Yanjun Ma, Juntao Li, Min Zhang
The paper introduces LENS, a lightweight algorithm that detects noisy labels in graph data by measuring how removing a node reduces neighborhood entropy. It shows that label noise increases entropy and blurs class boundaries, and proposes a relabeling step to preserve useful features. Experiments demonstrate that LENS outperforms existing noise‑robust methods in accuracy while being an order of magnitude faster.
By Pintu Kumar, Nandyala Hemachandra
The study evaluates large language models (LLMs) that generate political networks by comparing them to survey-reported partisan exposure from the ANES 2020–2022 Social Media Study. It finds that while network structure and political assortativity vary across model configurations, many models either over‑separate Democratic and Republican egos or underestimate Republican exposure, leading to group‑specific errors in local partisan environments. These discrepancies persist even among networks with similar aggregate properties, indicating that aggregate metrics alone can mask important individual‑level inaccuracies.
By Marco Tonin, Michele Tizzoni, Bruno Lepri, Luca Maria Aiello
MobileInject is a benchmark that evaluates the security of multimodal large language model (MLLM)-based mobile agents against multimodal injection attacks. It offers a high‑fidelity interactive framework with 11 scenarios, 9 risk categories, and 108 executable risk instances, and introduces Cross‑Modal Consistency Injection along with two advanced attack methods—Multi‑Role Collaborative Deception and Reasoning‑Camouflage Injection—to intensify trust manipulation and reasoning hijacking. Experiments show that mobile agents are broadly vulnerable, especially to early‑stage reasoning hijacking, establishing MobileInject as a systematic tool for studying and improving mobile agent security.
By Songping Wang, Peiyang Li, Yi Huang, Fengting Li, Xuhao Ran, Yueming Lyu, Chao Feng, Caifeng Shan
The paper introduces JointCoreBench, a benchmark for evaluating memory selection in large language model agents, and JointCore, a cost‑ordered group‑deletion baseline that identifies minimal sufficient memory subsets. JointCore achieves perfect recovery of minimum‑cost sufficient sets across 6,912 conditions, cuts subset evaluations by 62.9%, and saves 87.8% of active memory usage. Experiments across multiple model families and tasks demonstrate that JointCore preserves full‑memory behavior while reducing model queries and cost‑aware recovery rates compared to singleton deletion.
By BHVSP Subrahmanyam, Neeraj Choudhary