Tootfinder

Time-To-Inconsistency: A Survival Analysis of Large Language Model Robustness to Adversarial Attacks
Yubo Li, Ramayya Krishnan, Rema Padman
https://arxiv.org/abs/2510.02712 http…

Time-To-Inconsistency: A Survival Analysis of Large Language Model Robustness to Adversarial Attacks
Large Language Models (LLMs) have revolutionized conversational AI, yet their robustness in extended multi-turn dialogues remains poorly understood. Existing evaluation frameworks focus on static benchmarks and single-turn assessments, failing to capture the temporal dynamics of conversational degradation that characterize real-world interactions. In this work, we present the first comprehensive survival analysis of conversational AI robustness, analyzing 36,951 conversation turns across 9 stat…

CoDA: Agentic Systems for Collaborative Data Visualization
Zichen Chen, Jiefeng Chen, Sercan \"O. Arik, Misha Sra, Tomas Pfister, Jinsung Yoon
https://arxiv.org/abs/2510.03194

CoDA: Agentic Systems for Collaborative Data Visualization
Deep research has revolutionized data analysis, yet data scientists still devote substantial time to manually crafting visualizations, highlighting the need for robust automation from natural language queries. However, current systems struggle with complex datasets containing multiple files and iterative refinement. Existing approaches, including simple single- or multi-agent systems, often oversimplify the task, focusing on initial query parsing while failing to robustly manage data complexity…

IFDECORATOR: Wrapping Instruction Following Reinforcement Learning with Verifiable Rewards
Xu Guo, Tianyi Liang, Tong Jian, Xiaogui Yang, Ling-I Wu, Chenhui Li, Zhihui Lu, Qipeng Guo, Kai Chen
https://arxiv.org/abs/2508.04632

IFDECORATOR: Wrapping Instruction Following Reinforcement Learning with Verifiable Rewards
Reinforcement Learning with Verifiable Rewards (RLVR) improves instruction following capabilities of large language models (LLMs), but suffers from training inefficiency due to inadequate difficulty assessment. Moreover, RLVR is prone to over-optimization, where LLMs exploit verification shortcuts without aligning to the actual intent of user instructions. We introduce Instruction Following Decorator (IFDecorator}, a framework that wraps RLVR training into a robust and sample-efficient pipeline…

Jakobi Meyers ready to play Week 1 for Raiders following trade request: 'I asked, they said no' https://www.nfl.com/news/jakobi-meyers-ready-to-play-week-1-for-raiders-following-trade-request-i-asked-they-said-no…

Filing: VOA Director Michael Abramowitz has been fired after refusing what he called an "illegal" reassignment to run a broadcasting station in North Carolina (Scott Nover/Washington Post)
https://www.washingtonpost.com/business/2025/08/04/voa-abramowit…

VOA director fired after declining reassignment to low-level post
Kari Lake tried to dispatch Michael Abramowitz to North Carolina. When he refused, she fired him.

MagicGUI: A Foundational Mobile GUI Agent with Scalable Data Pipeline and Reinforcement Fine-tuning
Liujian Tang, Shaokang Dong, Yijia Huang, Minqi Xiang, Hongtao Ruan, Bin Wang, Shuo Li, Zhihui Cao, Hailiang Pang, Heng Kong, He Yang, Mingxu Chai, Zhilin Gao, Xingyu Liu, Yingnan Fu, Jiaming Liu, Tao Gui, Xuanjing Huang, Yu-Gang Jiang, Qi Zhang, Kang Wang, Yunke Zhang, Yuran Wang

MagicGUI: A Foundational Mobile GUI Agent with Scalable Data Pipeline and Reinforcement Fine-tuning
This paper presents MagicGUI, a foundational mobile GUI agent designed to address critical challenges in perception, grounding, and reasoning within real-world mobile GUI environments. The framework is underpinned by following six key components: (1) a comprehensive and accurate dataset, constructed via the scalable GUI Data Pipeline, which aggregates the largest and most diverse GUI-centric multimodal data to date from open-source repositories, automated crawling, and targeted manual annotatio…

From Wye-Delta to Cross-Square Recursion Configurations in Graphene-Based Quantum Hall Arrays
Ngoc Thanh Mai Tran, Marta Musso, Dominick S. Scaletta, Wei-Chen Lin, Valery Ortiz Jimenez, Dean G. Jarrett, Massimo Ortolano, Curt A. Richter, Chi-Te Liang, David B. Newell, Albert F. Rigosi
https://arxiv.org/abs/2508.03347

From Wye-Delta to Cross-Square Recursion Configurations in Graphene-Based Quantum Hall Arrays
In electrical metrology, the quantum Hall effect is accessed at the Landau level filling factor ν = 2 plateau to define and disseminate the unit of electrical resistance (ohm). The robustness of the plateau is only exhibited at this Landau level filling factor and thus places a constraint on the quantized resistances that are accessible when constructing quantized Hall array resistance standards (QHARS) using epitaxial graphene on SiC. To overcome devices constrained by using Hall elements in …

Modeling Emission-Line Surface Brightness in a Multiphase Galactic Wind: An O VI Case Study
Zirui Chen, Zixuan Peng, Kate H. R. Rubin, Timothy M. Heckman, Matthew J. Hayes, Yakov Faerman, Crystal L. Martin, S. Peng Oh, Drummond B. Fielding
https://arxiv.org/abs/2510.02443

Modeling Emission-Line Surface Brightness in a Multiphase Galactic Wind: An O VI Case Study
We present a fast and robust analytic framework for predicting surface brightness (SB) of emission lines in galactic winds as a function of radius up to $\sim 100$ kpc out in the circum-galactic medium. We model multi-phase structure in galactic winds by capturing emission from both the volume-filling hot phase (T $\sim 10^{6-7}$ K) and turbulent radiative mixing layers that host intermediate temperature gas at the boundaries of cold clouds (T $\sim 10^4$ K). Our multi-phase framework makes sig…

Holographic Gaseous Lenses for High-Power Lasers
Devdigvijay Singh, Ke Ou, Sida Cao, Victor M. Perez-Ramirez, Harsha Rajesh, Debolina Chakraborty, Caleb Redshaw, Pelin Dedeler, Albertine Oudin, Michelle M. Wang, Julia M. Mikhailova, Livia Lancia, Caterina Riconda, Pierre Michel, Matthew R. Edwards
https://arxiv.org/abs/2510.02659

Holographic Gaseous Lenses for High-Power Lasers
The capabilities of the world's highest energy and peak-power pulsed lasers are limited by optical damage, and further advances in high-intensity laser science will require optics that are substantially more robust than existing components. We describe here the experimental demonstration of off-axis diffractive gaseous lenses capable of withstanding extreme laser fluence and immune to cumulative damage. We used less than 8 mJ of energy from interfering ultraviolet laser pulses to holographicall…

Change Data Capture is a powerful technique enabling organisations to respond to real-time data changes. At Berlin Buzzwords, Muhammet Orazov explored FlinkCDC, a component of Apache Flink and discussed how it uses Flink's robust stream processing capabilities to provide a CDC pattern.
Watch the full session: https://