Results for “model”

Journal articles (OpenAlex) Nguyen Huy Hoang, Trần Xuân Hiệp, Le Thi Tam · WSEAS Transactions on Information Science and Applications archive Geopolitics

The U.S. - China Trade War and Trade Diversion in ASEAN: Evidence from ASEAN-U.S. Trade Flows and Policy Implications for Vietnam

This study examines how the U.S.-China trade war has affected trade diversion in the ASEAN Economic Community (AEC), with particular attention to Vietnam. Combining qualitative analysis with an extended gravity model estimated by Poisson Pseudo Maximum Likelihood (PPML), the paper uses ASEAN-U.S. trade data for 2015-2024 to assess whether higher U.S. tariffs on Chinese goods redirected trade toward ASEAN. The results indicate a positive and statistically significant…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Huu Khanh Linh Nguyen, Trang Bùi Thị Quỳnh, Trinh Nguyen Thi Hong et al. · Humanities and Social Sciences Communications Geopolitics

Coordinating environmental standards and trade competition: an adaptive coordination-transplant model for Vietnam

This article examines how legal systems can coordinate tensions between environmental standards and trade competition in the era of next-generation free trade agreements. The study is motivated by the growing gap between increasingly stringent external green trade requirements, especially in the European Union (EU), and the fragmented domestic legal architectures of developing countries such as Vietnam (European Commission, 2019; European Parliament and Council, 2023; UNDP…

Read at Journal articles (OpenAlex)

arXiv (research papers) Om Narayan SI

ChainWatch: A Kill Chain-Aligned Sequential Detection Framework for Multi-Step Attacks in MCP-Based AI Agent Systems

The Model Context Protocol (MCP) is an open-source standard that allows AI agents to connect to external tools, databases, and services. While this connectivity enables powerful agent capabilities, it also introduces multi-step attacks that existing per-call defenses cannot reliably detect. Attackers can compose individually benign tool invocations into malicious sequences that evade isolated inspection. This paper presents ChainWatch, a sequential detection framework for…

Read at arXiv (research papers)

arXiv (research papers) Jie Li SI

Trusted Credentials, Untrusted Behavior: Benchmarking LLM-Agent Security in High-Performance Computing

Large language model (LLM) agents are starting to take on routine work in high-performance computing (HPC), including monitoring Slurm jobs, diagnosing failed builds, inspecting simulation output, and coordinating scientific workflows. To do this work, an agent commonly acts under its user's credentials and inherits the user's access to files and the scheduler. This arrangement creates a failure mode that ordinary account-level controls do not capture. Adversarial…

Read at arXiv (research papers)

arXiv (research papers) Haoyan Luo SI

Persistent Sparse Autoencoders: Learning Feature-Specific Timescales in Language Model Representations

Sparse autoencoders (SAEs) decompose language model activations into sparse features, yet these models traditionally encode each token independently, failing to expose information that persists across a sequence. We first show that temporal persistence can naturally emerge in standard SAE features: after a feature activates, the hidden state remains aligned with its direction, and past activations help reconstruct later hidden states. How long this lasts varies widely across…

Read at arXiv (research papers)

Latinoamérica21 Ernesto Hernández Norzagaray Geopolitics

Perestroika without glasnost in Venezuela and Cuba

Venezuela and Cuba are experimenting with a limited economic opening to preserve political control, following the model of China and Vietnam without opening the door to democratic freedoms. O post Perestroika without glasnost in Venezuela and Cuba apareceu primeiro em Latinoamérica 21.

Read at Latinoamérica21

arXiv (research papers) Rodrigo Martins Boos SI

Local AI pre-screening for human triple-blind peer review in health sciences

Academic peer review is under mounting strain: NeurIPS 2025 received 21,575 submissions, ICLR 2025 received 11,603, and ICML 2025 received 12,107. This volume has outpaced the supply of qualified reviewers, and large language models (LLMs) are already filling the gap, largely undisclosed. An independent analysis of ICLR 2026 found roughly 21% of its 75,800 peer reviews were fully AI-generated, with over half showing some AI involvement (up from 15.8% in 2024). Documented…

Read at arXiv (research papers)

arXiv (research papers) Yu Cui SI

Refusal is Not Safety! Benchmarking Latent Safety Risks of LLM-Driven Content Humorization

Safety defenses for large language models (LLMs) have been extensively studied, with existing approaches focusing on attack detection and refusal mechanisms. Such fixed-form direct refusal strategies may introduce the risk of prefix injection attacks. Recent work has explored a new direction that leverages humor as an indirect refusal mechanism to mitigate over-refusal in jailbreak scenarios and reduce prefix injection risks. However, this approach implicitly assumes that…

Read at arXiv (research papers)

arXiv (research papers) Rabimba Karanjai SI

Context Contamination in LLM Analysis of Network Security Logs: Poison with Passive Prompt Injection and Mitigation Evaluation

Large Language Models are increasingly deployed in Security Operations Centers for log analysis tasks including summarization, alert triage, and threat investigation. These systems ingest logs from external-facing services and process network logs as natural language contexts to generate security insights. We demonstrate that this architectural pattern introduces a critical vulnerability: adversaries can embed prompt injection payloads in log-generating fields that persist…

Read at arXiv (research papers)

arXiv (research papers) Sanket Badhe SI

Agent Skill Security: Threat Models, Attacks, Defenses, and Evaluation

Reusable skills are becoming a fundamental building block of Large Language Model (LLM) agents, enabling capabilities to be packaged, shared, and reused across diverse applications. However, existing security research primarily focuses on prompt injection and runtime execution, leaving security risks throughout the broader skill lifecycle largely unexplored. In this paper, we present SkillSec-Eval, a lifecycle-aware framework for systematically evaluating the security of…

Read at arXiv (research papers)

arXiv (research papers) Junhui Wang SI

PVDetector: Detecting Prompt Injection Attacks on Purpose-Specific LLM Agents through Policy-Violation Concept Analysis

Large language models (LLMs) are increasingly deployed as purpose-specific agents to handle domain-specific tasks such as customer service and code generation. These agents are expected to comply with not only generic safety guardrails but also purpose-specific restrictions tailored to their designated roles. Such additional restrictions enlarge the attack surface, particularly to prompt injection (PI) attacks. To defend against such attacks, existing detection methods…

Read at arXiv (research papers)

arXiv (research papers) Huihao Jing SI

Isolation as a First-Class Principle for LLM-Agent System Safety: Concepts, Taxonomy, Challenges and Future Directions

The capability of LLM agents to function as the ``brain'' of a system fundamentally expands the scope of analysis beyond a standalone model. Consequently, safety is no longer only about input--output content alignment. It also concerns system behavior and real-world execution outcomes. However, the current literature is fragmented across attack types, applications, and benchmarks. This makes it hard to explain why failures such as prompt injection, tool misuse, and memory…

Read at arXiv (research papers)

Journal articles (OpenAlex) Qasir Shah · Journal of Philosophy of Education Geopolitics

When inclusion excludes: liberalism, Confucianism, and the Parkfield Community School protests

Abstract This article examines the 2019 Parkfield Community School protests in Birmingham as a case-based philosophical analysis in the moral and relational limits of liberal inclusion within public education. It critically engages Abeera Khan’s analysis of securitized liberalism, which shows how Muslim parental dissent was pathologized and rendered unintelligible within dominant civic discourses. At the same time, it draws on Nora Hämäläinen’s model of moral renegotiation…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Endre Jo Reite · Applied Economics Letters Geopolitics

Capital floors, model conservatism, and the limits of prudential incentives in green lending

Prudential green supporting factors and brown penalizing factors affect loan pricing only through the capital constraint that is active at the margin. Under Basel III, effective risk-weighted assets are the greater of internal-model estimates and the output-floor share of standardized totals. Standardized multipliers, therefore, affect marginal pricing directly only when the floor leg binds or when a policy change moves the bank across the floor threshold. Otherwise, pricing…

Read at Journal articles (OpenAlex)

arXiv (research papers) Ruksat Khan Shayoni SI

NetInjectBench: Benchmarking Indirect Prompt Injection in Tool-Using Large Language Model Agents for Network Operations

Tool-using large language model (LLM) agents are attractive for network operations, but tickets, alerts, logs, runbooks, and ChatOps messages can carry indirect prompt injections. We present NetInjectBench, a 130-scenario benchmark that separates untrusted artifact text, trusted policy metadata, and evaluation labels for network-operation tool use. The sample contains 40 benign, 40 weak-attack, 40 strong-attack, and 10 approved high-impact change scenarios; each is evaluated…

Read at arXiv (research papers)

arXiv (research papers) Yaxin Li SI

Devil in the Lens: Analyzing and Defending Physical Prompt Injection Against Vision-Language Models on Wearable Devices

Vision-Language Models (VLMs) are rapidly deployed on human-facing wearable devices such as smart glasses to enable multimodal perception and AI-assisted decision-making. While prior research has demonstrated the risks of visual prompt injection into digital image inputs of VLMs, the unique security challenges posed by the increasing integration between physical environments and wearable intelligence, such as those embodied in VLM-enabled AI glasses, remain underexplored…

Read at arXiv (research papers)

arXiv (research papers) Nguyen Khoi Vo SI

The Effect of Multi-Lingual and Keyword Adversarial Injection on LLM Relevance Judgment

Large language models (LLMs) are increasingly being used as automated judges for relevance evaluation in information retrieval, yet their robustness to adversarial manipulation remains insufficiently understood, particularly in multilingual settings. In this work, we investigate the impact of cross-lingual prompt injection attacks on LLM-based relevance judgments using TREC Deep Learning collections and two open-weight models under established prompting frameworks. We…

Read at arXiv (research papers)

arXiv (research papers) Mohammad Dabaja SI

Promptable Concept Segmentation from Above: Evaluating SAM 3's Zero-Shot and One-Shot Capabilities in Remote Sensing

The deployment of large-scale foundation models, such as the Segment Anything Model 3 (SAM 3), promises a transition toward open-vocabulary, training-free computer vision. However, their capacity to generalize out-of-distribution to the complex, top-down geometric structures of Earth Observation imagery remains largely unquantified. Driven by SAM 3's performance disparities in highly specialized domains, we present a comprehensive, multi-task empirical evaluation across…

Read at arXiv (research papers)

Journal articles (OpenAlex) Vo Thi Pho Chau, Minh Hiệp Ngô, M.A. Pham Van Thinh et al. · Small Wars and Insurgencies Geopolitics

The Strategic Hamlet Program in South Vietnam (1961–1965): pacification, rural control, and the collapse of the “Special War” strategy

This article re-evaluates the Strategic Hamlet Program (SHP) in South Vietnam (1961–1965), contending that its rapid disintegration stemmed from a fundamental crisis of legitimacy at the village level. Diverging from traditional scholarship that emphasizes administrative or military deficiencies, this study utilizes Vietnamese archival materials and regional case studies to illuminate the collision between an externally imposed counterinsurgency model and the autonomous…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Thi Thu Ha Vo, Chen Wu, Tiezhong Liu · Natural Hazards Review Geopolitics

Double-Hurdle Approach to Understanding Household Willingness to Pay for Public Flood Mitigation Infrastructure in Vietnam

Abstract The understanding of household willingness to pay for improving the public drainage infrastructure remains limited, especially in the low- and middle-income classes. To advance this literature, we construct a comprehensive behavioral framework that integrates the protection motivation theory, collective action theory, and social exchange theory to explain household financial engagement in urban flood mitigation. The double hurdle model is applied to disentangle the…

Read at Journal articles (OpenAlex)

arXiv (research papers) Hugo García Cuesta SI

Multi-Agent Firewall Architecture for Privacy Protection of Sensitive Data in Interactions with Language Models

While Large Language Models (LLMs) have become essential productivity tools, their integration into workflows without adequate safeguards creates significant risks. This paper proposes an open-source, privacy-focused, user-facing firewall designed to secure both web-based and programmatic LLM interactions. The architecture combines a browser extension and a proxy for total traffic interception across both HTTP(S) and WebSocket communications. At its core, a flexible…

Read at arXiv (research papers)

arXiv (research papers) Phat T. Tran-Truong SI

FedMark-FM: Auditable, Risk-Adjusted Data Markets for Federated Foundation-Model Adaptation

Federated foundation-model adaptation increasingly relies on heterogeneous private artifacts (retrieval corpora, prompts and demonstrations, LoRA adapters, preference and safety data, and update sketches), yet existing federated-learning incentive mechanisms price clients as homogeneous data or update providers. This assumption poorly matches foundation-model pipelines, where contribution value is heterogeneous, non-IID, pipeline-dependent, privacy-constrained, and…

Read at arXiv (research papers)

arXiv (research papers) Aya Spira SI

Beware of Agentic Botnets: Scalable Untargeted Promptware Attacks via Universal and Transferable Adversarial HalluSquatting

The growing adoption of agentic LLM applications has introduced a new threat previously named as promptware. While prior work has established that adversaries can exploit direct channels to LLM applications to apply promptware under weak threat models, many applications do not provide any direct channels that could be exploited for prompt injection beyond the Internet. This raises a question: can attackers exploit LLM applications at scale without any direct channels in…

Read at arXiv (research papers)

arXiv (research papers) Elad David SI

Latent Undertow: How Ordinary Typos Break Probes

LLMs handle ordinary typing variation fluently: a typo or missing punctuation leaves both user intent and the model's response substantively unchanged. Yet probes that detect malicious prompts by reading the model's hidden states tell a different story: the same edit rotates the readout vector by 43--56 at the perturbed token, decaying below 15% within ~10 downstream tokens. Stacking ~3 common typos per message cuts a single-position prompt-injection probe's TPR@FPR$=1% by…

Read at arXiv (research papers)

Article 19 natashas Geopolitics

Senegal: Parliament must be more transparent and strive for open governance

ARTICLE 19 Senegal and West Africa, together with parliamentary monitoring organisations (PMOs), are today calling for greater transparency, public participation and public accountability from Senegal’s parliament to make the institution a model of open governance. Senegal is ranked 24th out of 33 countries analysed in the African Open Parliament Index (AOPI). ‘Senegal’s current ranking in

Read at Article 19

Berkeley AI Research blog SI

Intelligence is Free, Now What? Data Systems for, of, and by Agents

... government of the people, by the people, for the people ... — Abraham Lincoln, Gettysburg Address (1863) The cost of AI is dropping rapidly. GPT-4-class capabilities cost roughly $30 per million tokens in early 2023; today the same runs under $1, and some providers are pushing costs below $0.10. Across benchmarks, inference prices have fallen between 9x and 900x per year, with a median decline near 50x. Even frontier models are getting dramatically cheaper each…

Read at Berkeley AI Research blog

Journal articles (OpenAlex) Nguyen Thi Minh Hanh · AAOU Journal/AAOU journal Geopolitics

Digital learning in Vietnamese prisons: an evidence-informed conceptual model for secure open and distance education in custodial settings

Purpose This article develops an evidence-informed conceptual model for secure open and distance education (ODE) in Vietnamese prisons. It treats prison digital learning as constrained-access ODE for incarcerated learners and people preparing for release and examines its potential contribution to upskilling, reskilling and reintegration under custodial constraints. Design/methodology/approach The article uses secondary evidence synthesis, combining legal-policy mapping…

Read at Journal articles (OpenAlex)

arXiv (research papers) Bogdan Banu SI

Biological Motifs for Agentic Control

The transition of Large Language Models (LLMs) from passive generators to autonomous agents has introduced significant challenges in reliability, security, and state management. Current agentic architectures are often constructed ad-hoc, prone to hallucination cascades, infinite loops, and prompt injection attacks. This paper argues that many of these failure modes can be analyzed using control motifs long studied in systems biology, provided the comparison is made at the…

Read at arXiv (research papers)

arXiv (research papers) Amit LeVi SI

Unsupervised Features Mining via Activation Geometry

Interpretability methods aim to reveal the features represented inside large language models (LLMs). Many existing methods begin with labeled examples of a human-defined concept that may reflect human biases, and then identify how that concept is represented within the model, for example in its activation space or through other decomposition methods. We introduce \emph{Mining via Activation Geometry} (MAG), a simple unsupervised framework for extracting reasoning features…

Read at arXiv (research papers)

Older stories