Tootfinder

Generalizing Verifiable Instruction Following
Valentina Pyatkin, Saumya Malik, Victoria Graf, Hamish Ivison, Shengyi Huang, Pradeep Dasigi, Nathan Lambert, Hannaneh Hajishirzi
https://arxiv.org/abs/2507.02833

Generalizing Verifiable Instruction Following
A crucial factor for successful human and AI interaction is the ability of language models or chatbots to follow human instructions precisely. A common feature of instructions are output constraints like ``only answer with yes or no" or ``mention the word `abrakadabra' at least 3 times" that the user adds to craft a more useful answer. Even today's strongest models struggle with fulfilling such constraints. We find that most models strongly overfit on a small set of verifiable constraints from …

A Hierarchical and Evolvable Benchmark for Fine-Grained Code Instruction Following with Multi-Turn Feedback
Guoliang Duan, Mingwei Liu, Yanlin Wang, Chong Wang, Xin Peng, Zibin Zheng
https://arxiv.org/abs/2507.00699

A Hierarchical and Evolvable Benchmark for Fine-Grained Code Instruction Following with Multi-Turn Feedback
Large language models (LLMs) have advanced significantly in code generation, yet their ability to follow complex programming instructions with layered and diverse constraints remains underexplored. Existing benchmarks often prioritize functional correctness, overlooking the nuanced requirements found in real-world development. We introduce MultiCodeIF, a comprehensive benchmark designed to evaluate instruction-following in code generation across multiple dimensions: constraint type, hierarchica…

SEC filing: Super Micro cautions that its financial control weaknesses, if not fixed, could hurt its ability to report results "in a timely and accurate manner" (Dina Bass/Bloomberg)
https://www.bloomberg.com/news/articles/20

Is the hyperscaling relation violated below the upper critical dimension in some particular cases?
Hung T. Diep, Van-Thanh Ngo
https://arxiv.org/abs/2507.02159

Is the hyperscaling relation violated below the upper critical dimension in some particular cases?
In this review, we show our results with new interpretation on the critical exponents of thin films obtained by high-performance multi-histogram Monte Carlo simulations. The film thickness $N_z$ consists of a few layers up to a dozen of layers in the $z$ direction. The free boundary condition is applied in this direction while in the $xy$ plane periodic boundary conditions are used. Large $xy$ plane sizes are used for finite-size scaling. The Ising model is studied with nearest-neighbor (NN) in…

CIMR: Contextualized Iterative Multimodal Reasoning for Robust Instruction Following in LVLMs
Yangshu Yuan, Heng Chen, Xinyi Jiang, Christian Ng, Kexin Qiu
https://arxiv.org/abs/2507.22074

CIMR: Contextualized Iterative Multimodal Reasoning for Robust Instruction Following in LVLMs
The rapid advancement of Large Language Models (LLMs) and Large Vision-Language Models (LVLMs) has enhanced our ability to process and generate human language and visual information. However, these models often struggle with complex, multi-step multi-modal instructions that require logical reasoning, dynamic feedback integration, and iterative self-correction. To address this, we propose CIMR: Contextualized Iterative Multimodal Reasoning, a novel framework that introduces a context-aware itera…

Group Relative Policy Optimization for Speech Recognition
Prashanth Gurunath Shivakumar, Yile Gu, Ankur Gandhe, Ivan Bulyko
https://arxiv.org/abs/2509.01939 https://

Group Relative Policy Optimization for Speech Recognition
Speech Recognition has seen a dramatic shift towards adopting Large Language Models (LLMs). This shift is partly driven by good scalability properties demonstrated by LLMs, ability to leverage large amounts of labelled, unlabelled speech and text data, streaming capabilities with auto-regressive framework and multi-tasking with instruction following characteristics of LLMs. However, simple next-token prediction objective, typically employed with LLMs, have certain limitations in performance and…

The DOGE team at SSA might have violated FISMA and other laws by not following security protocols as spelled out in NIST's SP 800-53, which are mandatory for all government agencies.
It's no surprise then that a whistleblower is warning that we have lost the ability to see who is accessing 300 million Americans' most sensitive information after DOGE moved SSA data to their own Amazon cloud instance.
Thanks to John Skinner, former project lead for 18F, for his expert i…

Whistleblower: DOGE put Social Security database covering 300 million Americans on insecure cloud
The complaint accuses DOGE of bypassing security protocols to move Americans’ sensitive personal data outside federal oversight and onto an insecure AWS cloud instance, potentially violating established security practices and several laws.

Non-local random deposition models for earthquakes and energy propagation
Philippe Carmona, Fran\c{c}ois P\'etr\'elis, Nicolas P\'etr\'elis
https://arxiv.org/abs/2506.23853

Non-local random deposition models for earthquakes and energy propagation
We investigate a new class of non-local random deposition models, initially introduced by physicists to study the field of mechanical constraints (stress) applied along a line or on a given area located in a seismic zone. The non-local features are twofold. First, the falling objects have random and heavy-tailed dimensions. Second, the locations where the objects are falling are at least for some of the models that we consider, depending on the shape of the surface before deposition. We conside…

A Survey of Scientific Large Language Models: From Data Foundations to Agent Frontiers
Ming Hu, Chenglong Ma, Wei Li, Wanghan Xu, Jiamin Wu, Jucheng Hu, Tianbin Li, Guohang Zhuang, Jiaqi Liu, Yingzhou Lu, Ying Chen, Chaoyang Zhang, Cheng Tan, Jie Ying, Guocheng Wu, Shujian Gao, Pengcheng Chen, Jiashi Lin, Haitao Wu, Lulu Chen, Fengxiang Wang, Yuanyuan Zhang, Xiangyu Zhao, Feilong Tang, Encheng Su, Junzhi Ning, Xinyao Liu, Ye Du, Changkai Ji, Cheng Tang, Huihui Xu, Ziyang Chen, Ziyan Hu…

A Survey of Scientific Large Language Models: From Data Foundations to Agent Frontiers
Scientific Large Language Models (Sci-LLMs) are transforming how knowledge is represented, integrated, and applied in scientific research, yet their progress is shaped by the complex nature of scientific data. This survey presents a comprehensive, data-centric synthesis that reframes the development of Sci-LLMs as a co-evolution between models and their underlying data substrate. We formulate a unified taxonomy of scientific data and a hierarchical model of scientific knowledge, emphasizing the…

Kraken secures a MiCA license from the Central Bank of Ireland to offer crypto services across the EU, following Coinbase's approval in Luxembourg last week (Callan Quinn/Decrypt)
https://decrypt.co/327112/kraken-license-europes-new-rules

Kraken Becomes Latest to Win a License Under Europe's New Rules
Approval from Ireland's central bank gives Kraken the ability to operate across 27 EU states as it races to catch up to rival Coinbase.