Carlos's Debrief

⚡ Quick Summary

A Red-Team Study of Anthropic Fable 5 & Opus 4.8 Models: We evaluate the adversarial robustness of two frontier large language models (LLMs) developed by Anthropic, Fable 5 and Opus 4.8, against four families of automated jailbreak attack across 7 826 harmful intents spanning a ten-category harm…
Evaluating Open-Source LLMs for Multi-Label ATT&CK Technique Classification on CTI Reports: Classifying Cyber Threat Intelligence (CTI) using MITRE Adversarial Tactics, Techniques, and Common Knowledge (ATT&CK) is essential for proactive defense, but historically required extensive human effort. Pre-Large Language Model (LLM)…
Verifiable computations for dynamic encrypted control: Encrypted control can preserve the privacy of data and parameters while the necessary computations can be outsourced to a cloud server. To ensure the integrity of the received values from the cloud, i.e., that they have not been changed,…
Learning Red Agent Policy from Observations for Neurosymbolic Autonomous Cyber Agents: With sophisticated cyber-attacks becoming increasingly prevalent, modern networks require intelligent autonomous cyber-defense agents trained via Reinforcement Learning (RL). These agents employ neurosymbolic approaches such as behavior…
Learning from the Self-future: On-policy Self-distillation for dLLMs: On-policy self-distillation (OPSD) has proven effective for post-training large language models (LLMs), yet its application to diffusion LLMs (dLLMs) remains unexplored. Existing OPSD methods are inherently autoregressive-centric.

📰 News Headlines

LoopCoder-v2: Only Loop Once for Efficient Test-Time Computation Scaling: Looped Transformers scale latent computation by repeatedly applying shared blocks, but sequential looping increases latency and KV-cache memory with the loop count.
ACE-Ego-0: Unifying Egocentric Human and Robotic Data for VLA Pretraining: Vision-Language-Action (VLA) models benefit from large-scale and diverse embodied data, yet scaling robot trajectory collection is costly and labor-intensive.
GameCraft-Bench: Can Agents Build Playable Games End-to-End in a Real Game Engine?: Game generation is an emerging application of coding agents, requiring models to transform natural-language specifications into playable interactive systems.
LectūraAgents: A Multi-Agent Framework for Adaptive Personalized AI-Assisted Learning and Embodied Teaching: Effective personalized AI-assisted learning demands systems that can not only generate accurate learner-specific educational materials, but also dynamically adapt their instruction to diverse…
TRIAGE: Dialectical Reasoning for Explainable Risk Prediction on Irregularly Sampled Medical Time Series with LLMs: Clinical early warning systems built on electronic health records, in which clinical observations are recorded as irregularly sampled medical time series (ISMTS), must deliver both calibrated risk…

🗡️ Relevant to your Katana work 3

📄 ArXiv Papers

🧠 Rethinking / Dataset / Distillation 2

Learning from the Self-future: On-policy Self-distillation for dLLMs [46]

Yifu Luo, Zeyu Chen, Haoyu Wang, Xinhao Hu, Yuxuan Zhang, Zhizhou Sha, Shiwei Liu

On-policy self-distillation (OPSD) has proven effective for post-training large language models (LLMs), yet its application to diffusion LLMs (dLLMs) remains unexplored. Existing OPSD methods are…

cs.CL

Rethinking Dataset Distillation for Classification: Do Distilled Sets Outperform Coresets? [14]

Trisha Mittal, Akshay Mehra, Joshua Kimball

Dataset distillation (DD) has emerged as a prominent approach in data centric machine learning, aiming to synthesize compact training sets for efficient training by compressing the information in…

cs.LG

🧠 Att / Multi-source / Cybersecurity 2

Evaluating Open-Source LLMs for Multi-Label ATT&CK Technique Classification on CTI Reports [28]

Ahmed Ryan, Saad Sakib Noor, Md Erfan, Shaswata Mitra, Sudip Mittal, Md Rayhanur Rahman

Classifying Cyber Threat Intelligence (CTI) using MITRE Adversarial Tactics, Techniques, and Common Knowledge (ATT&CK) is essential for proactive defense, but historically required extensive human…

cs.CRcs.LG

Multi-Source Cybersecurity Logs: An ATT&CK-Labeled Dataset and SLM Evaluation [16]

Abir Ashab Niloy, Ahmed Ryan, Imamul Hossain Rafi, Md Erfan, Md Rayhanur Rahman

Multi-stage cyberattacks span system, network, and browser logs. Detecting them requires correlating events across all three sources. Machine learning methods can learn these cross-source patterns,…

cs.CRcs.LG

🧠 Visual / Verification / Enables 1

Visual Verification Enables Inference-time Steering and Autonomous Policy Improvement [1]

Mingtong Zhang, Dhruv Shah

Robots deployed in the real world should learn from their experience and improve over time. This requires a mechanism of practicing and learning from feedback. In this paper, we propose VERITAS, a…

cs.ROcs.AI

🧠 Reprorepo / Scaling / Reproducibility 1

ReproRepo: Scaling Reproducibility Audits with GitHub Repository Issues [2]

Shanda Li, Qiuhong Anna Wei, Jingwu Tang, Valerie Chen, Nihar B Shah, Tim Dettmers, Yiming Yang, Ameet Talwalkar

Reproducing research results from papers and released code is central to scientific progress. Existing works have introduced benchmarks to evaluate whether LLM agents can assist with reproducibility,…

cs.CLcs.AIcs.LG

🧠 Evolvenav / Proactive / Preflection 1

EvolveNav: Proactive Preflection and Self-Evolving Memory for Zero-Shot Object Goal Navigation [3]

Qi Chai, Wenhao Shen, Nanjie Yao, Yue Xia, Kaiyong Zhao, Jie Ma, Guosheng Lin, Hao Wang

Zero-Shot Object-Goal Navigation (ZS-OGN) requires embodied agents to explore and locate target objects without any prior training. To this end, recent methods leverage foundation models. But they…

cs.AI

🧠 Red / Agent / Policy 1

Learning Red Agent Policy from Observations for Neurosymbolic Autonomous Cyber Agents [4]

Ankita Samaddar, Sandeep Neema, Daniel Balasubramanian, Xenofon Koutsoukos

With sophisticated cyber-attacks becoming increasingly prevalent, modern networks require intelligent autonomous cyber-defense agents trained via Reinforcement Learning (RL). These agents employ…

cs.CRcs.AIcs.LG

🧠 Looped / World 1

Looped World Models [5]

Hongyuan Adam Lu, Z. L. Victor Wei, Qun Zhang, Jinrui Zeng, Bowen Cao, Lingwei Meng, Mocheng Li, Zezhong Wang, Haonan Yin, Naifu Xue, Minyu Chen, Cenyuan Zhang, Zefan Zhang, Hao Wei, Jiawei Zhou, Haoran Xu, Hao Yang, Ronglai Zuo, Tongda Xu, Yonghao Li, Jian Chen, Hebin Wang, Zeyu Gao, Yang Li, Wei Zhao, Qimin Zhong, Siqi Liu, Yumeng Zhang, Leyan Cui, Zhangyu Wang, Wai Lam

Current world models face a fundamental tension: faithful long-horizon simulation demands deep computation, but deeper models are expensive to deploy and prone to compounding errors. We resolve this…

cs.LGcs.AIcs.CL

🧠 Fixed-point / Reasoners / Stable 1

Fixed-Point Reasoners: Stable and Adaptive Deep Looped Transformers [6]

Sajad Movahedi, Vera Milovanović, Shlomo Libo Feigin, Alexander Theus, Thomas Hofmann, Valentina Boeva, T. Konstantin Rusch, Antonio Orvieto

Looped architectures provide an inductive bias toward learning step-by-step procedures for tasks that require compositional reasoning. The number of effective layers reached by looping determines the…

cs.AI

🧠 Health / Rubricstree / Scalable 1

RubricsTree: Scalable and Evolving Open-Ended Evaluation of Personal Health Agents across Health Memory and Medical Skills [7]

Weizhi Zhang, Zechen Li, Hamid Palangi, Ben Graef, A. Ali Heydari, Simon A. Lee, Salman Rahman, Ray Luo, Zeinab Esmaeilpour, Erik Schenck, Chloe Zhang, Yamin Li, Menglian Zhou, Philip S. Yu, Daniel McDuff, Lindsey Sunden, Mark Malhotra, Shwetak Patel, Ahmed A. Metwally

The LLM-empowered personal health agents with user health (sensor) metrics have offered a promising pathway to alleviate global disparities in healthcare access. However, large-scale clinical…

cs.CLcs.AI

🧠 Red-team / Anthropic / Fable 1

A Red-Team Study of Anthropic Fable 5 & Opus 4.8 Models [8]

Nicola Franco

We evaluate the adversarial robustness of two frontier large language models (LLMs) developed by Anthropic, Fable 5 and Opus 4.8, against four families of automated jailbreak attack across 7 826…

cs.CRcs.AIcs.CL

🧠 Stanford / Edgar / Filings 1

The Stanford EDGAR Filings Dataset: Reconstructing U.S. Corporate and Financial Disclosures into Layout-Faithful and Token-Efficient Pretraining Data [9]

Nick Bettencourt, Xiaowei Ding, Kay Giesecke

As high-quality public web corpora become increasingly exhausted, clean long-context documents have become a scarce and expensive source of training data for large language models (LLMs). Existing…

cs.AI

🧠 Drflow / Research / Benchmark 1

DRFLOW: A Deep Research Benchmark for Personalized Workflow Prediction [10]

Md Tawkat Islam Khondaker, Raymond Li, Muhammad Abdul-Mageed, Laks V. S. Lakshmanan, Issam H. Laradji

Deep research (DR) systems are increasingly used for complex information-seeking tasks, but existing works mainly focus on generating reports and summaries. In contrast, many enterprise tasks instead…

cs.AIcs.MA

🧠 Sign-rank / Index / List 1

Sign-Rank, Index, and List Replicability: Connections and Separations [11]

Ari Blondal, Hamed Hatami, Pooya Hatami, Chavdar Lalov, Sivan Tretiak

In learning theory, the sign rank of a binary concept class captures the smallest dimension in which it can be represented by points and halfspaces. Despite tremendous interest, lower bounds on sign…

cs.LGcs.IT

🧠 Adaptive / Volumetric / Mechanical 1

Adaptive Volumetric Mechanical Property Fields Invariant to Resolution [12]

Rishit Dagli, Donglai Xiang, Vismay Modi, Xuning Yang, Gavriel State, David I. W. Levin, Maria Shugrina

Accurate mechanical properties (or materials) Young's modulus ($E$), Poisson's ratio ($ν$) and density ($ρ$) are essential for reliable physics simulation of digital worlds, but most 3D assets lack…

cs.CVcs.LGcs.RO

🧠 Finite-time / Queue / Peak 1

Finite-Time Queue Peak Laws in Stochastic Networks: Logarithmic Scaling After Geometric Thresholds [13]

Hao Liang, Cheng Tang, Yunzong Xu

We study finite-horizon queue peaks in generalized switches, a standard stochastic-network model in which many queues share constrained service resources. Arrivals may be dependent, time-varying, and…

math.PRcs.LGeess.SY

🧠 Conformal / Prediction / Intervals 1

Conformal Prediction Intervals with Tail-Specific Guarantees [15]

Simone Cuonzo, Nina Deliu

This paper extends classical conformal frameworks for constructing prediction intervals with global marginal coverage $1-α$ to intervals that provide explicitly calibrated guarantees for the upper…

math.STq-fin.RMstat.ME

🧠 Kolmogorov / Regression / Robust 1

Kolmogorov Regression for Robust Diffusion Policies [17]

Lekan Molu

Finite-dimensional (FD) diffusion policies exhibit temporal drift owing to discretization artifacts that degrade long-horizon performance (when deployed on physical systems). We introduce a backward…

cs.LGcs.AI

🧠 Superconductivity / Interband / Coupling 1

Superconductivity from interband coupling to ferroelectric quantum critical fluctuations in two dimensions [18]

Sudip Kumar Saha, Jonathan Ruhman, Avraham Klein

Soft critical fluctuations associated with ferroelectric quantum phase transitions are typically transverse owing to their polar nature. This implies that the conventional density--density…

cond-mat.supr-concond-mat.str-el

🧠 Future / Dynamic / Reconstruction 1

Future Dynamic 3D Reconstruction: A 3D World Model with Disentangled Ego-Motion [19]

Nils Morbitzer, Jonathan Evers, Artem Savkin, Thomas Stauner, Nassir Navab, Federico Tombari, Stefano Gasperini

Forecasting the evolution of dynamic environments is crucial for autonomous agents. While generative world models have recently achieved high photorealism in 2D video synthesis by mixing ego-motion…

cs.CV

🧠 Unified / Multimodal / Autoregressive 1

Unified Multimodal Autoregressive Modeling with Shared Context-Visual Tokenizer is Key to Unification [20]

Wujian Peng, Lingchen Meng, Yuxuan Cai, Xianwei Zhuang, Yuhuan Yang, Rongyao Fang, Chenfei Wu, Junyang Lin, Zuxuan Wu, Shuai Bai

Unified Multimodal Modeling aims to integrate visual understanding and generation within a single system. However, existing approaches typically rely on two disparate visual tokenizers, which splits…

cs.CV

🧠 Variable-width / Transformers 1

Variable-Width Transformers [21]

Zhaofeng Wu, Oliver Sieberling, Shawn Tan, Rameswar Panda, Yury Polyanskiy, Yoon Kim

Scaling model size, specifically depth and width, has driven significant progress in transformer-based language models. However, most architectures maintain a constant width across all layers,…

cs.CL

🧠 Pahspecs / Spatially / Resolved 1

PAHSPECS: Spatially Resolved PAH Spectroscopy at cosmic noon with JWST MIRI MRS [22]

Fergus R. Donnan, Karin Sandstrom, Irene Shivaei, Leindert A. Boogaard, Tanio Díaz-Santos, Cristina M. Lofaro, Román Fernández Aranda, Manuel Aravena, Roberto Decarli, Hanae Inami, Pablo G. Pérez-González, Gergö Popping, Paul P. van der Werf

We present spatially resolved spectroscopy with JWST/MIRI MRS of a representative sample of normal star-forming galaxies at $z\sim1.1$ as part of the PAHSPECS program. To extract emission from…

astro-ph.GA

🧠 Mochi / Motion / Enhancement 1

MOCHI: Motion Enhancement of Collaborative Human-object Interactions [23]

Jiye Lee, Yonghun Choi, Jungdam Won

Collaborative human-object interaction shows dynamic and complex movements that require mutual anticipation and continuous adjustment between participants and the shared object. Modeling such…

cs.CVcs.GRcs.RO

🧠 Eventdrive / Event / Cameras 1

EventDrive: Event Cameras for Vision-Language Driving Intelligence [24]

Dongyue Lu, Rong Li, Ao Liang, Lingdong Kong, Wei Yin, Lai Xing Ng, Benoit R. Cottereau, Camille Simon Chane, Wei Tsang Ooi

Event cameras sense the world through asynchronous brightness changes with microsecond latency and high dynamic range, offering motion fidelity far beyond frame-based sensors and capturing temporal…

cs.CV

🧠 Joint / Optimal / Search 1

A Joint Optimal Search for Gravitational Waves from Resolved and Unresolved Supermassive Binary Black Holes with Pulsar Timing Arrays [25]

Boris Goncharov, Gabriela Sato-Polito, Xiaoming Bi, Matias Zaldarriaga

We introduce, from first principles, a joint model of the gravitational wave background (GWB) and brightest supermassive black hole binary (SMBHB) sources that may be individually resolvable in…

astro-ph.HEastro-ph.IMgr-qc

🧠 Gatling / Rapid-fire / Consensus 1

Gatling: Rapid-Fire Consensus from Parallel Composition [26]

Giulia Scaffino, Max Resnick, Joachim Neu

Consensus protocols form the core of blockchains and other replicated state machines, ensuring that all correct nodes process the same totally ordered log of input transactions. In fault-free…

cs.CRcs.DC

🧠 Seeing / Not / Screening 1

Seeing Is Not Screening: Multimodal Hidden Instruction Attacks on Agent Skill Scanners [27]

Xiaojun Jia, Jie Liao, Simeng Qin, Ke Ma, Wenbo Guo, Yebo Feng, Aishan Liu, Yang Liu

Agent skills are emerging as an important attack surface in LLM-based systems. Through an empirical study of existing skill scanners, we find that current defenses primarily rely on textual…

cs.CRcs.CV

🧠 Structural / Role / Injection 1

Structural Role Injection in Handlebars-Templated LLM Prompts: Triple-Brace Interpolation, Delimiter Family, and the Limits of HTML Auto-Escaping [29]

Mohammadreza Rashidi

Large language model applications build prompts from templates, and Handlebars is a widely used templating engine and the default prompt-template format in Microsoft Semantic Kernel. Its double-brace…

cs.CRcs.AIcs.CL

🧠 Verifiable / Computations / Dynamic 1

Verifiable computations for dynamic encrypted control [30]

Sebastian Schlor, Frank Allgöwer

Encrypted control can preserve the privacy of data and parameters while the necessary computations can be outsourced to a cloud server. To ensure the integrity of the received values from the cloud,…

eess.SYcs.CR

🧠 Structural / Preservation / Logical 1

Structural Preservation and the Logical Expressiveness of Graph Neural Networks [31]

Przemysław Andrzej Wałęga, Bernardo Cuenca Grau

Bridges between graph neural networks (GNNs) and logical formalisms have been established by fixing architectural choices, such as the types of aggregation, combination, and activation functions.…

cs.AI

🧠 Koopman / Lifting / Certified 1

Koopman Lifting with Certified Error Bounds for Joint Inference in Nonlinear Networks [32]

Chuansen Peng, Xiaojing Shen, Yunmin Zhu

Jointly inferring latent node states and unknown network topology in nonlinear graphical dynamical systems is a fundamental yet largely unsolved problem, where the mutual entanglement of continuous…

math.OC

🧠 Flowclip / Contrastive / Pretraining 1

FlowCLIP: Contrastive Pretraining Using Domain Names for Encrypted Traffic Classification [33]

Eun Hun Choi

Network traffic classification enables website fingerprinting, intrusion detection, and Quality of Service management. However, developing methods that capture stable and generalizable traffic…

cs.NI

🧠 Period / Finding / Lattice 1

From Period Finding to Lattice Sampling: Experimental Insights into Shor's and Regev's Factoring Algorithms [34]

Daniela Falcó, Arturo Rodríguez, Guillermo Rivas, Ricardo S. Alonso

Quantum algorithms for integer factorization represent one of the most prominent applications of quantum computation, with far-reaching implications for modern cryptography. While Shor's algorithm…

quant-ph

🧠 Zero-sum / Problems / Two 1

On zero-sum problems of two new types [35]

Zhi-Wei Sun

In this paper, we mainly investigate zero-sum problems over $\mathbb Z/n\mathbb Z$ (with $n>1$) of two new types. Let $s_1(n)$ (resp. $t_1(n)$) be the least positive integer $k$ such that for any…

math.NTmath.CO

🧠 Graph / Darshana / Parallel 1

Darshana Graph: A Parallel Commentary Corpus for Comparative Indian Philosophy, with Stylometric and Exploratory Graph Analyses [36]

Joy Bose

We introduce Darshana Graph, a corpus of over 125,000 text records spanning classical Hindu, Buddhist, and Jain philosophical traditions, drawn from public-domain and openly licensed translations of…

cs.CLcs.DL

🧠 Zone / Proximal / Policy 1

Zone of Proximal Policy Optimization: Teacher in Prompts, Not Gradients [37]

Byung-Kwan Lee, Ximing Lu, Shizhe Diao, Minki Kang, Saurav Muralidharan, Karan Sapra, Andrew Tao, Pavlo Molchanov, Yejin Choi, Yu-Chiang Frank Wang, Ryo Hachiuma

Knowledge distillation transfers a teacher's competence to a small student but is brittle in the small-student regime: forcing the student to imitate logits from a much larger teacher concentrates it…

cs.CL

🧠 Impulse / Decoding / Quantum 1

Impulse Decoding of Quantum LDPC Codes: Equivalence of Degeneracy and Code-Shortening [38]

Shobhit Bhatnagar, Michele Pacenti, Nithin Raveendran, David Declercq, Bane Vasić

Quantum error correction is essential for building scalable quantum computers. Within the stabilizer formalism, the Calderbank-Shor-Steane framework constructs quantum codes from pairs of classical…

quant-ph

🧠 Twist-3 / Contributions / Factorization 1

Twist-3 contributions to $γγ\toπ^0π^0,\,K_S^0K_S^0$ in $k_T$ factorization [39]

Jun-Kang He, Cong Wang

We compute the cross sections for the two-photon processes $γγ\toπ^0π^0$ and $γγ\to K_S^0K_S^0$ in $k_T$ factorization, including the chirally enhanced two-parton twist-3 light-cone distribution…

hep-ph

🧠 Quartic / Lyapunov / Functions 1

Quartic Lyapunov functions for global fluid stability [40]

David Darrow, Elizabeth Carlson, David Goluskin

A fluid system is 'globally stable' if all initial conditions eventually converge to the same state. Since Reynolds (1895) and Orr (1907), the standard way to show global stability has been the…

physics.flu-dyn

🧠 Strategies / Preventing / Reversing 1

Strategies for preventing and reversing polarized online discourse [41]

Leon Klingborg, Kenneth Mavor, Alexander J. Stewart

Political polarization poses a variety of challenges for modern democracies. Entrenched disagreements on policy can prevent constructive discourse and compromise, and high levels of affective…

physics.soc-phcs.CY

🧠 Directed / Reachability-preserving / Minimum 1

Directed Reachability-Preserving Minimum Edge Cut: Approximation and Planar Hardness [42]

Qi Duan

We study a directed version of the three-terminal reachability-preserving minimum edge cut problem. Given a directed graph $G=(V,A)$ with arc costs and terminals $s_1,s_2,t$, the one-way directed…

cs.DScs.CC

🧠 Benchmark / Suite / Intracellular 1

A benchmark suite of intracellular Boolean model variants and multiscale simulations for computational biology [43]

Marco Masera, Riccardo Smeriglio, Roberta Bardini, Alessandro Savino, Stefano Di Carlo

We present PhysiBench, an open resource for developing and evaluating computational methods in systems biology including a benchmark suite of 612 executable intracellular Boolean regulatory network…

cs.DB

🧠 Extreme / Ram-pressure / Stripping 1

An extreme ram-pressure stripping event in a protocluster at redshift 4.3 [44]

Dazhi Zhou, Scott C. Chapman, Manuel Aravena, Roger Deane, Anthony H. Gonzalez, Ryley Hill, Nicholas LeVar, Matthew A. Malkan, Adam Muzzin, Nan Zhang, Kedar A. Phadke, Vismaya R. Pillai, Manuel Solimano, Justin S. Spilker, Nikolaus Sulzenauer, Joaquin D. Vieira, David Vizgan, George C. P. Wang, Axel Weiss

In the nearby Universe, the environment plays a crucial role in suppressing star formation in dense regions. In particular, ram-pressure stripping (RPS) is a major mechanism for removing gas from…

astro-ph.GA

🧠 Analyzing / Encoding / Al-mawrid 1

Analyzing and Encoding the Al-Mawrid Arabic-English Dictionary with the ISO Language Markup Framework and TEI Lex-0 [45]

Diaa Fayed, Laurent Romary

This paper presents a robust methodology for the systematic digitization and encoding of the Al-Mawrid Arabic-English dictionary, transforming it from a legacy print resource into a standardized…

cs.CL

🌐 Web Findings

🤗 HuggingFace Papers 16

Aligning Quantum Operators with Large Language Models [67]

Can Large Language Models (LLMs) understand and reason about quantum operators? Despite their remarkable capabilities in mathematics and symbolic reasoning, LLMs remain inherently…

▲ 1HuggingFace Papers

A Gradient Perspective on RLVR Stability and Winner Advantage Policy Optimization [62]

Reinforcement learning with verifiable rewards (RLVR) improves language-model reasoning, but GRPO-style optimization remains prone to collapse. We analyse this instability through…

▲ 6HuggingFace Papers

TRIAGE: Dialectical Reasoning for Explainable Risk Prediction on Irregularly Sampled Medical Time Series with LLMs [65]

Clinical early warning systems built on electronic health records, in which clinical observations are recorded as irregularly sampled medical time series (ISMTS), must deliver…

▲ 24HuggingFace Papers

Beyond Monolingual Deep Research: Evaluating Agents and Retrievers with Cross-Lingual BrowseComp-Plus [61]

Deep research agents are increasingly evaluated on their ability to search for evidence, reason over retrieved sources, and produce grounded answers. Existing browsing benchmarks,…

▲ 1HuggingFace Papers

ChLogic: Evaluating Robustness of Logical Reasoning in Chinese Expressions [54]

Large language models perform increasingly well on standardized logical reasoning benchmarks, but whether this ability remains robust beyond English is unclear. We introduce…

▲ 6HuggingFace Papers

ProCUA-SFT Technical Report [58]

Training computer-use agents (CUAs) -- models that interact with graphical desktops through screenshots and keyboard/mouse actions -- requires large-scale, diverse trajectory data…

▲ 3HuggingFace Papers

ACE-Ego-0: Unifying Egocentric Human and Robotic Data for VLA Pretraining [57]

Vision-Language-Action (VLA) models benefit from large-scale and diverse embodied data, yet scaling robot trajectory collection is costly and labor-intensive. Recent advances show…

▲ 34HuggingFace Papers

Rethinking the Role of Efficient Attention in Hybrid Architectures [53]

Modern language models increasingly adopt hybrid architectures that combine full attention with efficient attention modules, such as sliding-window attention (SWA) and recurrent…

▲ 7HuggingFace Papers

GameCraft-Bench: Can Agents Build Playable Games End-to-End in a Real Game Engine? [63]

Game generation is an emerging application of coding agents, requiring models to transform natural-language specifications into playable interactive systems. Unlike traditional…

▲ 34HuggingFace Papers

OPD-Evolver: Cultivating Holistic Agent Evolver via On-Policy Distillation [60]

Memory has become a standard substrate for self-evolving agents, yet retaining experience is not the same as learning how to evolve through it. Existing memory agents can store…

▲ 20HuggingFace Papers

ActWorld: From Explorable to Interactive World Model via Action-Aware Memory [64]

Interactive world models aim to simulate environment dynamics under real-time user actions. However, their action vocabulary is largely confined to navigation: most actions…

▲ 3HuggingFace Papers

LectūraAgents: A Multi-Agent Framework for Adaptive Personalized AI-Assisted Learning and Embodied Teaching [66]

Effective personalized AI-assisted learning demands systems that can not only generate accurate learner-specific educational materials, but also dynamically adapt their…

▲ 25HuggingFace Papers

MotionVLA: Vision-Language-Action Model for Humanoid Motion [55]

Generating realistic humanoid motion from scene images and text involves both low-frequency pose semantics and high-frequency physical dynamics. However, many existing methods…

HuggingFace Papers

Show the Signal, Hide the Noise: Spectral Forcing for Pixel-Space Diffusion [56]

Pixel-space diffusion models are trained on full-bandwidth noisy images, yet the useful signal available to the denoiser is strongly frequency dependent. Under rectified-flow…

▲ 14HuggingFace Papers

Text-Vision Co-Instructed Image Editing [52]

Existing image editing methods can be generally categorized into textual instruction-based and visual prompt-based ones. Textual instructions are semantically expressive, but are…

HuggingFace Papers

LoopCoder-v2: Only Loop Once for Efficient Test-Time Computation Scaling [59]

Looped Transformers scale latent computation by repeatedly applying shared blocks, but sequential looping increases latency and KV-cache memory with the loop count. Parallel loop…

▲ 58HuggingFace Papers

📝 OpenReview 2

SVD-LLM: Truncation-aware Singular Value Decomposition for Large Language Model Compression [77]

The advancements in Large Language Models (LLMs) have been hindered by their substantial sizes, which necessitate LLM compression methods for practical deployment. Singular Value…

OpenReview

Diffusion$^2$: Dynamic 3D Content Generation via Score Composition of Video and Multi-view Diffusion Models [78]

Recent advancements in 3D generation are predominantly propelled by improvements in 3D-aware image diffusion models. These models are pretrained on Internet-scale image data and…

OpenReview

🦞 Lobste.rs 5

Honda Civics and the Evil Valet [50]

Honda Civics and the Evil Valet Status update, three years later June 13, 2026 · 6 min · Eric McDonald Table of Contents Keys to the Kingdom Building Tools Outstanding Work - A…

Lobste.rs

I Could've Rickrolled the Entire FIFA World Cup. All I Needed Was My ID [47]

How I found that anyone could register on FIFA

Lobste.rs

A 27-Year-Old Authentication Bypass in OpenBSD's PPP Stack [48]

OpenBSD's sppp_pap_input function used attacker-controlled length fields as the bcmp comparison length for credential validation. Sending zero-length name and password fields…

Lobste.rs

The Impact of Google's Manifest Version 3 Update on Ad Blocker Effectiveness [51]

Google's recent update to the manifest file for Chrome browser extensions, transitioning from manifest version 2 (MV2) to manifest version 3 (MV3), has raised concerns among users…

Lobste.rs

Humiliating IIS servers for fun and jail time [49]

A friend of mine once told me: If you ever spot an IIS blue screen, don’t stop there; there must be something.

Lobste.rs