The study audits AI "humanizers"—tools that rewrite machine‑generated text to evade detection—by checking whether factual units such as numbers, dates, names, and quoted spans survive each rewrite. Using a rule‑based checker validated by AI annotators, the authors found that in the 2026 public leaderboard, 35–65% of documents had at least one factual unit altered or dropped, with quoted spans most affected. An open 12B humanizer retained most numbers and names but reworded 67–78% of quotes, and a zero‑shot detector (Jev) could distinguish human originals from humanized rewrites with an AUROC of 0.864, outperforming several existing detectors.
By Ziyuan Wang, Genliang Zhu, Xiaobo Chen
ShiftRAG is a new retrieval‑augmented generation framework that decouples discriminative retrieval learning from autoregressive generation while still using a single LLM backbone. It addresses the textual bottleneck of existing RAG pipelines by projecting retrieval alignments into continuous soft tokens, creating a high‑bandwidth interface that feeds retrieval signals into the generator. The approach also introduces a soft orthogonality objective to reduce rank collapse and layer‑wise relevance dynamics to preserve fine‑grained latent space, resulting in significant improvements in retrieval and question‑answering performance over state‑of‑the‑art baselines.
By Youmin Ko, Jihong Jeong, Hyunjoon Kim
IdeaTaste is a benchmark that evaluates large language models (LLMs) on their ability to generate research proposals that align with what human researchers actually publish. The study compares proposals generated by six contemporary LLMs against real papers from four research areas, finding that LLMs achieve only about 26% of the citation matching and 18% of the idea matching of human communities. Even larger LLM populations and richer field contexts do not close this gap, indicating that LLMs struggle more with translating research directions into concrete approaches than with selecting the right topics.
By Shanshan Zhong, Chen Xu, Jingjie Ning, Zhengzhong Liu, Aviral Kumar, Chenyan Xiong
Skill++ addresses lifelong skill routing for large language model agents by rewriting the name and description of each new skill instead of retraining the router. This approach preserves earlier skills while allowing new ones to be selected, achieving 76.3% query success versus 36.4% with original metadata across a library of 731 skills. It outperforms existing methods on three benchmarks and five LLM routers, and its benefits transfer to the Codex CLI agent harness.
By Hanrong Zhang, Minghao Liang, Jiawei Wu, Yankai Chen, Shicheng Fan, Zhiwei Liu, Enze Ma, Junwei Yuan, Tianyi Wang, Dehai Min, Kening Zheng, Xue Liu, Philip S. Yu
The paper introduces a comprehensive audit protocol for System One decision models, which provide typed, probability‑scored answers without human oversight. Using three adapted benchmarks—BBQ, CrowS‑Pairs, and SilverSpoon—it evaluates twelve models plus a no‑thinking baseline across 1,638,000 decisions, revealing widespread stereotype alignment, order‑dependent responses, and confidence instability. The study demonstrates that swapping answer options and reversing question order significantly alters model outputs, underscoring the need for order‑invariant scoring in procurement.
By Alejandro Mozo Quesada (Universidad de M\'alaga, Spain), Sergio G\'alvez Rojas (Universidad de M\'alaga, Spain), Francisco L\'opez Valverde (Universidad de M\'alaga, Spain)
The paper investigates whether large language model (LLM) coding agents can function as forward‑deployed engineer (FDE) agents by discovering customer requirements and executing deployment plans in a real enterprise text‑to‑SQL scenario. In a simulated engagement, agents must uncover acceptance criteria, data permissions, and hardware constraints before planning and delivering a post‑training recipe. Across 15 models, the best agents uncovered over 80% of hidden requirements, yet only 25% of runs met all acceptance criteria, revealing bottlenecks in discovery, execution reliability, compute usage, and benchmark relevance.
By Marianne Menglin Liu, Rongguang Wang, Zheng Wang, Yewen Wang, Leonid Boytsov, Tao Sheng, Sujith Ravi, Dan Roth
XDecideBench is a benchmark comprising nine suites across 122 languages, containing 88,425 texts and 153,875 typed questions, designed to evaluate Typed Decision Models and their confidence policies. The benchmark reports results separately for aligned, locally labelled, and native items, revealing that model rankings vary by suite and language, with the best English model performing poorly in many other languages. Fine‑tuned open models outperform their base Qwen counterparts in all twelve development languages, and confidence thresholds set on English data can exceed acceptable error rates on other datasets such as XNLI.
By Marek \v{S}uppa, Jaroslav Kop\v{c}an
WebHarbor is a collection of 153 dockerized website environments that faithfully replicate real-world web services, offering 3,518 tasks that include multi-step and post-authentication workflows. Built through task-driven human-agent collaboration, the suite ensures deterministic control, visual fidelity, and functional correctness. The platform enables reliable evaluation and cost-efficient reinforcement learning, with a Qwen 4B-level GUI agent trained in WebHarbor generalizing well to unseen tasks and matching leading open-source agents on four benchmarks.
By Zhaoyang Wang, Qianhui Wu, Boyu Gou, Demi Wang, Jiacheng Zhang, Peichun Hua, Qize Li, Shi Qiu, Yifan Jiang, Yifan Zhang, Zexu Jin, Baolin Peng, Bolin Shen, Chetan Bansal, Dehan Kong, Deren Lei, Dongshen Peng, Fengqing Jiang, Hao Cheng, Haoqing Wang, Jiaming Shan, Jicheng Wang, Jingguang Zhou, Jixuan He, Kaide Zeng, Kangyu Zhu, Lama Moukheiber, Mufan Qiu, Peng Xia, Qihan Sun, Qingwei Lin, Qiushi Sun, Ruichen Zhang, Shuqing Luo, Si Qin, Sicheng Fan, Tao Xiong, Wenlin Yao, Xinyu Zhang, Xinyue Zeng, Xuanrui Li, Xuchao Zhang, Yian Wang, Yichao Liu, Yiming Liang, Yiyang Zhou, Yizhu Jiao, Yuetai Li, Yurun Chen, Yuxuan Zhang, Zheyuan Yang, Zhiyuan Yao, Zhongyang Li, Zihao Li, Zili Wang, Jianfeng Gao, Huaxiu Yao
The paper titled "Tug of Watts: A study on the Energy - Accuracy Trade-off of Large Language Models on Medical Question Answering Tasks" examines how to estimate energy consumption and carbon emissions for 13 large language models (ranging from 270 M to 14 B parameters) used in medical QA. It highlights that reporting energy and accuracy separately is insufficient, advocating for joint trade‑off metrics to guide model selection. The authors aim to help practitioners choose more efficient models and build sustainable LLM pipelines.
By Viju Sudhi, Giulio Billi, Andreea Scrob, Regina Stodden, Philipp Cimiano
The paper presents a lightweight YOLOv11n-based framework for real‑time recognition of Nigerian Sign Language alphanumeric gestures on low‑resource mobile devices. It introduces a benchmark dataset of 11,208 images across 33 static classes, with a training set of 23,509 images and separate validation and test sets of 1,681 images each. The model, optimized with quantization‑aware compression and a spatial‑density occlusion index, achieves high precision (0.9975) and recall (0.9991) with inference latencies between 405.5 ms and 1,759.4 ms on Android hardware.
By Adeleke Adeniyan, Abimbola H. Afolayan, Olanrewaju V. Johnson
CoworkOS is a general‑purpose harness designed for long‑horizon real‑world automation, separating task strategy from execution management. Its three‑layer architecture allows direct execution or delegation to subagents, while a shared application layer lets humans and agents collaboratively edit documents, designs, and 3D scenes. The system incorporates hierarchical workspaces and persistent semantic memory to maintain continuity across agents and sessions, and demonstrates strong performance across 12 benchmarks and six application categories, outperforming leading agents like Codex and DeepSeek Harness.
By Jiajie Jin, Guanting Dong, Chengjun Pan, Xin Deng, Zhuohui Sheng, Guoqiang Wang, Tao Gui, Yutao Zhu, Yujia Chen, Shihong Yi, Binze Hu, Zhixiong Yang, Zhichao Sun, Rui Wang, Han Han, Mingyu Ma, Junhao Huang, Xiaotao Wang, Huicheng Jiang, Yuyang Hu, Xinyu Yang, Wanghan Xu, Yajie Yang, Zhecan Wang, Honglin Guo, Xiaoyu Xing, Xiaoqian Zhang, Yulun Zhu, Xiaoxi Li, Jiahang Lin, Shaofan Liu, Shihan Dou, Zhiheng Xi, Qi Zhang, Yu-Gang Jiang, See-Kiong Ng, Yang Zhang, Tat-Seng Chua, Ji-Rong Wen, Zhicheng Dou
The paper introduces ARCBench, a configurable benchmark for measuring rule compliance in language‑model agents that compact long histories to fit context windows. Using ARCBench, the authors find that rule violations after compaction rarely stem from rule deletion; instead, agents often ignore the rule while it remains present in the compacted history. The study shows that the source of violations depends on how the rule is delivered and on the compaction strategy, with targeted compactor attacks significantly increasing violations, and that simply retaining a rule in the compacted summary does not guarantee compliance.
By Yiwei Zha, Qingyu Meng, Herbert Bos, Min Chen
The paper introduces EmoNet‑Face‑HQ, a benchmark of 2,500 AI‑generated portraits each evaluated by four psychology experts across 40 emotion categories. It shows that most vision‑language models (VLMs) perform no better than random guessing unless fine‑tuned, but that changing the prompt to a simple yes/no per category can raise a VLM’s agreement with experts from chance to near expert levels. When evaluated using probability‑of‑yes scores on the five most‑agreed‑upon categories, eleven open‑weight VLMs achieve a quadratic‑weighted κ of 0.425, below the 0.468 two‑expert agreement, yet their average probability‑of‑yes score (0.553) matches the benchmark’s fine‑tuned model and outperforms it in three cases without any training.
By Tobias Hallmen, Fabian Deuser, Robin-Nico Kampa, Norbert Oswald, Elisabeth Andr\'e
The paper introduces hinted scaling laws, which incorporate privileged information to transform a benchmark into a continuum of easier versions, ensuring even small models achieve non‑zero accuracy. This approach yields smoother, more predictable accuracy trends, allowing accurate extrapolation to large models even when standard scaling laws fail. Across multiple pretraining ladders and model families, hinted scaling laws dramatically reduce forecasting errors and the number of test examples needed to distinguish model performance.
By Suze van Adrichem, Luke Bailey, Rohith Kuditipudi, Christine Ye, Neil Band, Ahmed Ahmed, John Duchi, Percy Liang, Tatsunori Hashimoto
ViSSFlow is a unified flow‑matching framework that jointly tackles Video‑to‑Sound (V2S) and Visual Text‑to‑Speech (VisualTTS) generation. It incorporates a Diffusion Transformer with a disentangled condition aggregation mechanism, using cross‑attention for semantic cues and self‑attention for dense visual information. The authors also introduce a feature‑level data synthesis strategy to compensate for limited joint training data, and demonstrate that a single model can be trained end‑to‑end without sacrificing performance on either task.
By Xin Cheng, Yuyue Wang, Xihua Wang, Yihan Wu, Kaisi Guan, Yijing Chen, Peng Zhang, Xiaojiang Liu, Meng Cao, Ruihua Song
The paper introduces PRISM, a framework that separates belief estimation from action selection to handle irregular state transitions in online medical consultations. By maintaining a posterior belief over consultation states, PRISM improves clinical decision accuracy by 22.40 percentage points and response quality by 5.51 BERTScore_F1 points compared to leading LLMs and clinical agents. Ablation studies confirm that oracle state information further enhances response quality across baselines.
By Mingzhe Lu, Xiangru Chen, Jian Qin, Yunpeng Li, Xihan Gao
The paper investigates attribution graphs for sparse autoencoders (SAEs) by examining how zeroing a feature affects subsequent layer activations. It shows that in TopK SAEs, removing a feature often causes another feature to cross the activation cutoff, and these crossings account for most of the response’s energy. The authors demonstrate that a margin‑based graph, which adds first‑order changes to preactivations, achieves low error (≤0.033) while accurately capturing the impact of feature ablations.
By Karl Li
arXiv:2611.00460v1 Announce Type: cross
Abstract: Large language models (LLMs) now answer medical exam questions at or above expert level, but it is unclear whether such scores reflect clinical compe...
By Navapat Nananukul, Mayank Kejriwal
arXiv:2611.00081v1 Announce Type: cross
Abstract: Digital libraries preserve mathematical books printed from the sixteenth to the twentieth century, but the lettered diagrams in them, the figures on...
By Minh Quang Tran, Carter He, Cameron Bloch, Zitai Huang, Thien Truong Vo, Trong Nguyen Tran, Isaac Chen
arXiv:2611.02338v1 Announce Type: new
Abstract: AI assistants store values that a language model derived from a person's records under the person's own definitions, such as their important contacts....
By Zichuan Xu, Haochen Zhang, Yuyang Dong, Makoto Onizuka, Chuan Xiao