The growing adoption of large language models (LLMs) has led to a new paradigm in mobile computing--LLM-powered mobile AI agents--capable of decomposing and automating complex tasks directly on smartphones. However, the security implications of these agents remain largely unexplored. In this paper, we present the first comprehensive security analysis of mobile LLM agents, encompassing three representative categories: System-level AI Agents developed by original equipment manufacturers (e.g., YOYO Assistant), Third-party Universal Agents (e.g., Zhipu AI AutoGLM), and Emerging Agent Frameworks (e.g., Alibaba Mobile Agent). We begin by analyzing the general workflow of mobile agents and identifying security threats across three core capability dimensions: language-based reasoning, GUI-based interaction, and system-level execution. Our analysis reveals 11 distinct attack surfaces, all rooted in the unique capabilities and interaction patterns of mobile LLM agents, and spanning their entire operational lifecycle. To investigate these threats in practice, we introduce AgentScan, a semi-automated security analysis framework that systematically evaluates mobile LLM agents across all 11 attack scenarios. Applying AgentScan to nine widely deployed agents, we uncover a concerning trend: every agent is vulnerable to targeted attacks. In the most severe cases, agents exhibit vulnerabilities across eight distinct attack vectors. These attacks can cause behavioral deviations, privacy leakage, or even full execution hijacking. Based on these findings, we propose a set of defensive design principles and practical recommendations for building secure mobile LLM agents. Our disclosures have received positive feedback from two major device vendors. Overall, this work highlights the urgent need for standardized security practices in the fast-evolving landscape of LLM-driven mobile automation.

In Text-to-SQL, execution feedback is essential for guiding large language models (LLMs) to reason accurately and generate reliable SQL queries. However, existing methods treat execution feedback solely as a post-hoc signal for correction or selection, failing to integrate it into the generation process. This limitation hinders their ability to address reasoning errors as they occur, ultimately reducing query accuracy and robustness. To address this issue, we propose ReEx-SQL (Reasoning with Execution-Aware Reinforcement Learning), a framework for Text-to-SQL that enables models to interact with the database during decoding and dynamically adjust their reasoning based on execution feedback. ReEx-SQL introduces an execution-aware reasoning paradigm that interleaves intermediate SQL execution into reasoning paths, facilitating context-sensitive revisions. It achieves this through structured prompts with markup tags and a stepwise rollout strategy that integrates execution feedback into each stage of generation. To supervise policy learning, we develop a composite reward function that includes an exploration reward, explicitly encouraging effective database interaction. Additionally, ReEx-SQL adopts a tree-based decoding strategy to support exploratory reasoning, enabling dynamic expansion of alternative reasoning paths. Notably, ReEx-SQL achieves 88.8% on Spider and 64.9% on BIRD at the 7B scale, surpassing the standard reasoning baseline by 2.7% and 2.6%, respectively. It also shows robustness, achieving 85.2% on Spider-Realistic with leading performance. In addition, its tree-structured decoding improves efficiency and performance over linear decoding, reducing inference time by 51.9% on the BIRD development set.

In light of the rapid adoption of AI coding assistants, LLM-assisted development has become increasingly prevalent, creating an urgent need for robust evaluation of generated code quality. Existing benchmarks often require extensive manual effort to create static datasets, rely on indirect or insufficiently challenging tasks, depend on non-scalable ground truth, or neglect critical low-level security evaluations, particularly memory-safety issues. In this work, we introduce OSS-Bench, a benchmark generator that automatically constructs large-scale, live evaluation tasks from real-world open-source software. OSS-Bench replaces functions with LLM-generated code and evaluates them using three natural metrics: compilability, functional correctness, and memory safety, leveraging robust signals like compilation failures, test-suite violations, and sanitizer alerts as ground truth. In our evaluation, the benchmark, instantiated as OSS-Bench(php) and OSS-Bench(sql), profiles 17 diverse LLMs, revealing insights such as intra-family behavioral patterns and inconsistencies between model size and performance. Our results demonstrate that OSS-Bench mitigates overfitting by leveraging the evolving complexity of OSS and highlights LLMs' limited understanding of low-level code security via extended fuzzing experiments. Overall, OSS-Bench offers a practical and scalable framework for benchmarking the real-world coding capabilities of LLMs.

Identifying subtle technical errors within complex scientific and technical documents, especially those requiring multimodal interpretation (e.g., formulas in images), presents a significant hurdle for Large Language Models (LLMs) whose inherent error-correction tendencies can mask inaccuracies. This exploratory proof-of-concept (PoC) study investigates structured LLM context conditioning, informed by Persistent Workflow Prompting (PWP) principles, as a methodological strategy to modulate this LLM behavior at inference time. The approach is designed to enhance the reliability of readily available, general-purpose LLMs (specifically Gemini 2.5 Pro and ChatGPT Plus o3) for precise validation tasks, crucially relying only on their standard chat interfaces without API access or model modifications. To explore this methodology, we focused on validating chemical formulas within a single, complex test paper with known textual and image-based errors. Several prompting strategies were evaluated: while basic prompts proved unreliable, an approach adapting PWP structures to rigorously condition the LLM's analytical mindset appeared to improve textual error identification with both models. Notably, this method also guided Gemini 2.5 Pro to repeatedly identify a subtle image-based formula error previously overlooked during manual review, a task where ChatGPT Plus o3 failed in our tests. These preliminary findings highlight specific LLM operational modes that impede detail-oriented validation and suggest that PWP-informed context conditioning offers a promising and highly accessible technique for developing more robust LLM-driven analytical workflows, particularly for tasks requiring meticulous error detection in scientific and technical documents. Extensive validation beyond this limited PoC is necessary to ascertain broader applicability.Keywords: AI-assisted, AI-powered, AI-enhanced, automated, knowledge engineering, machine learning.

Although large language models (LLMs) perform well in general tasks, domain-specific applications suffer from hallucinations and accuracy limitations. Continual Pre-Training (CPT) approaches encounter two key issues: (1) domain-biased data degrades general language skills, and (2) improper corpus-mixture ratios limit effective adaptation. To address these, we propose a novel framework, Mixture of Losses (MoL), which decouples optimization objectives for domain-specific and general corpora. Specifically, cross-entropy (CE) loss is applied to domain-corpus to ensure knowledge acquisition, while Kullback-Leibler (KL) divergence aligns general-corpus training with the base model's foundational capabilities. This dual-loss architecture preserves universal skills while enhancing domain expertise, avoiding catastrophic forgetting. Empirically, we validate that a 1:1 domain-to-general corpus ratio optimally balances training and overfitting without the need for extensive tuning or resource-intensive experiments. Furthermore, our experiments demonstrate significant performance gains compared to traditional CPT approaches, which often suffer from degradation in general language capabilities; our model achieves 27.9% higher accuracy on the Math-500 benchmark in the non-think reasoning mode, and an impressive 83.3% improvement on the challenging AIME25 subset in the think mode, underscoring the effectiveness of our approach.

Demis Hassabis, CEO of Google DeepMind, says that reaching artificial general intelligence or AGI--a fuzzy term typically used to describe machines with human-like cleverness--will mean honing some of the nascent abilities found in Google's flagship Gemini models. Google announced a slew of AI upgrades and new products at its annual I/O event today in Mountain View, California. The search giant revealed upgraded versions of Gemini Flash and Gemini Pro, Google's fastest and most capable models, respectively. Hassabis said that Gemini Pro outscores other models on LMArena, a widely used benchmark for measuring the abilities of AI models. Hassabis showed off some experimental AI offerings that reflect a vision for artificial intelligence that goes far beyond the chat window.

Good Job is Slate's advice column on work. Have a workplace problem big or small? I am a hard-line hater of generative AI (ChatGPT, Midjourney, etc.). I think it's bad for the environment and bad for society. It burns water resources, exploits workers in the global south, plagiarizes art and writing, and eliminates badly needed entry-level jobs.

Despite the fact that billions of dollars are being poured into reshaping energy infrastructure around the needs of AI, no one has settled on a way to quantify AI's energy usage. Worse, companies are generally unwilling to disclose their own piece of the puzzle. There are also limitations to estimating the emissions associated with that energy demand, because the grid hosts a complicated, ever-changing mix of energy sources. So, that said, here are the many variables, assumptions, and caveats that we used to calculate the consequences of an AI query. Companies like OpenAI, dealing in "closed-source" models, generally offer access to their systems through an interface where you input a question and receive an answer.

Recent advances in large language models (LLMs) have fueled the vision of automated scientific discovery, often called AI Co-Scientists. To date, prior work casts these systems as generative co-authors responsible for crafting hypotheses, synthesizing code, or drafting manuscripts. In this work, we explore a complementary application: using LLMs as verifiers to automate the \textbf{academic verification of scientific manuscripts}. To that end, we introduce SPOT, a dataset of 83 published papers paired with 91 errors significant enough to prompt errata or retraction, cross-validated with actual authors and human annotators. Evaluating state-of-the-art LLMs on SPOT, we find that none surpasses 21.1\% recall or 6.1\% precision (o3 achieves the best scores, with all others near zero). Furthermore, confidence estimates are uniformly low, and across eight independent runs, models rarely rediscover the same errors, undermining their reliability. Finally, qualitative analysis with domain experts reveals that even the strongest models make mistakes resembling student-level misconceptions derived from misunderstandings. These findings highlight the substantial gap between current LLM capabilities and the requirements for dependable AI-assisted academic verification.

Reward modeling is essential for aligning large language models with human preferences through reinforcement learning from human feedback. To provide accurate reward signals, a reward model (RM) should stimulate deep thinking and conduct interpretable reasoning before assigning a score or a judgment. Inspired by recent advances of long chain-of-thought on reasoning-intensive tasks, we hypothesize and validate that integrating reasoning capabilities into reward modeling significantly enhances RMs interpretability and performance. To this end, we introduce a new class of generative reward models - Reasoning Reward Models (ReasRMs) - which formulate reward modeling as a reasoning task. We propose a reasoning-oriented training pipeline and train a family of ReasRMs, RM-R1. RM-R1 features a chain-of-rubrics (CoR) mechanism - self-generating sample-level chat rubrics or math/code solutions, and evaluating candidate responses against them. The training of RM-R1 consists of two key stages: (1) distillation of high-quality reasoning chains and (2) reinforcement learning with verifiable rewards. Empirically, our models achieve state-of-the-art performance across three reward model benchmarks on average, outperforming much larger open-weight models (e.g., INF-ORM-Llama3.1-70B) and proprietary ones (e.g., GPT-4o) by up to 4.9%. Beyond final performance, we perform thorough empirical analyses to understand the key ingredients of successful ReasRM training. To facilitate future research, we release six REASRM models along with code and data at https://github.com/RM-R1-UIUC/RM-R1.