Results for “Again”

arXiv (research papers) Lin Sun SI

External Experience Serving in Production LLM Systems: A Deployment-Oriented Study of Quality-Cost Trade-offs

Production LLM systems accumulate reusable operational experience, but the practical deployment issue is not merely whether such experience can help. It is how different serving strategies trade off quality against online cost under realistic constraints. Injecting external experience can improve task quality, yet it also increases prompt burden, latency, and serving pressure. We study \textit{external experience serving} as a deployment-oriented quality-cost trade-off…

Read at arXiv (research papers)

The Guardian Socialism Norman Solomon

The spectre of gen Z socialism is haunting the world … according to the Economist | Norman Solomon

The magazine writes: ‘Resisting Gen-Z socialism is therefore an urgent task.’ That urgency must outweigh any urgency of feeding hungry people A spectre is haunting Europe and America – the spectre of gen Z socialism. That’s the urgent warning from the Economist in a new cover-story editorial, How to fight back against Gen-Z socialism. Alarmed by a youthful threat to the established order, the magazine is calling for heightened vigilance from defenders of private enterprise…

Read at The Guardian Socialism

arXiv (research papers) David Hofer SI

Assessing Automated Prompt Injection Attacks in Agentic Environments

Indirect prompt injection poses a critical threat to LLM agents that interact with untrusted external data, yet automated attack methods--proven effective for jailbreaking--remain underexplored in realistic agentic settings. We present a comprehensive empirical evaluation of automated prompt injection attacks against LLM agents, adapting both white-box (GCG) and black-box (TAP) methods to the agentic setting within the AgentDojo framework. We evaluate across 80 task pairs…

Read at arXiv (research papers)

arXiv (research papers) Hyunseok Paeng SI

The Injection Paradox: Brand-Level Suppression in Safety-Trained LLM Recommendations via RAG Context Injection

We present a reproducible failure mode of safety training in RAG-based LLM recommendation, the Injection Paradox, in which prompt injections embedded in retrieved documents backfire against the attacker, suppressing the target brand below the injection-free baseline. In safety-trained Claude models, documents containing prompt injections suffer a sharp drop in recommendation rate, and this suppression propagates beyond the injected document to unmodified documents of the…

Read at arXiv (research papers)

arXiv (research papers) Wenbo Guo SI

MalSkillBench: A Runtime-Verified Benchmark of Malicious Agent Skills

AI coding agents such as Claude Code and Gemini CLI increasingly extend themselves with third-party skills: markdown packages bundling natural-language instructions, executable scripts, and tool permissions. Because a skill is at once code and agent-facing instruction, it introduces a supply chain dependency whose risk is neither pure code nor pure prompt. Detection tools have never been measured against verified ground truth spanning this hybrid space, leaving their…

Read at arXiv (research papers)

The Conversation Geopolitics

Why is Australia buying used submarines? A naval expert answers key AUKUS questions

Following news of Australia acquiring used Virginia-class submarines from the US, AUKUS has been in the headlines again, but how much of what you’ve heard is true?

Read at The Conversation

Journal articles (OpenAlex) Aube Tollu · International Affairs Geopolitics

The social and relational life of Islamic State: a queer lens on gender, love and community-making under the caliphate

Abstract This article harnesses a queer lens to explore the intimate and relational infrastructure of French members of Islamic State (IS) through first-hand interviews conducted in France. It calls for an empirically grounded queer scholarship in International Relations that challenges heteronormative framings of armed actors, demonstrating the potential of queer theory to study non-sexual orientation and gender identity-diverse populations. Against the drawbacks of the…

Read at Journal articles (OpenAlex)

arXiv (research papers) Nicholas Saban SI

Domain-Conditioned Safety in Frontier Computer-Using Agents: A 793-Episode Browser Benchmark, a Coding-Domain Cross-Reference, and a Reproducibility Audit of Recent Red-Teaming

Recent computer-using-agent (CUA) red-teaming papers report prompt-injection attack success rates (ASR) of 42-98%, but these headline numbers cluster on retired models and on the most-vulnerable model in each paper's panel. We ask whether those techniques, reproduced as hand-crafted templates, still work against current frontier CUAs. We release CUA-HandCrafted, a public benchmark of 793 episodes spanning 24 multi-step web tasks, 56 attack templates, 8 attack families, and 4…

Read at arXiv (research papers)

Journal articles (OpenAlex) Jelena Pia-Comella · Current Sociology Geopolitics

The synergies between the Geneva Conventions and the Convention on the Elimination of All Forms of Discrimination against Women to combat sexual and gender-based violence against women in both times of peace and armed conflict

This article reviews feminist literature to contextualize how sexual and gender-based violence against women is rooted in discriminatory gender-power dynamics. Indeed, sexual and gender-based violence against women, such as rape, forced marriage and lack of access to sexual and reproductive health care, constitutes a gross violation of women’s rights in times of peace, which is exacerbated during armed conflicts and can amount to war crimes. The article will then offer an…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Yuree Noh, Fatimah Saadi · Politics & Gender Geopolitics

The Patriarchal Peace: Violence Against Women in Politics in Authoritarian Contexts

Abstract What does violence against women in politics (VAWIP) look like in authoritarian regimes that are widely portrayed as stable and “peaceful”? Existing scholarship largely centers on democracies, leaving nondemocratic contexts understudied. This article addresses the gap by examining Kuwait, the Gulf’s most democratic monarchy, where women’s educational and professional gains are highlighted even as political contestation remains tightly controlled. Drawing on 35…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Yuanyuan Cui · English Language Teaching Geopolitics

Multilingual Practices in China-Vietnam Borderlands: A Case of Supermarket at Hekou

Against the backdrop of increasingly frequent cross-border mobility, language plays an important role in catering to the needs of foreign migrants who come to China for various purposes. This study selected a supermarket in Hekou, a China-Vietnam border town, as a research site for examining how different languages are displayed across different areas of the supermarket and what functions these languages serve. Based on the content analysis of 127 multilingual signs and…

Read at Journal articles (OpenAlex)

arXiv (research papers) Amr Sobhy Geopolitics

Conflict or Strategy? Asymmetric Role Framing of La France insoumise and Rassemblement National in French News Headlines, 2022-2025

Do French news headlines frame left- and right-populist challengers as symmetric ``extremes,'' or as fundamentally different political adversaries? We examine 28,592 headlines about La France insoumise (LFI) and Rassemblement National (RN) published by 25 French-language outlets between 2022 and 2025, annotated through a three-model LLM pipeline validated against a stratified human audit. The clearest finding is role asymmetry rather than valence asymmetry: conflict framing…

Read at arXiv (research papers)

arXiv (research papers) Jiejun Tan SI

From Prompt Injection to Persistent Control: Defending Agentic Harness Against Trojan Backdoors

LLM agents are evolving from conversational chatbots to operational tools in real-world workspaces. In local agentic harnesses, an LLM can read and write files, call tools, and reuse workspace state across sessions. While such capabilities enhance utility, they also expose a new attack surface for attackers. Attackers can embed a prompt injection within a file or tool output. Agents may read this hidden instruction, store it, and execute it later. In this multi-step trojan…

Read at arXiv (research papers)

arXiv (research papers) Junyoung Park SI

Persona Attack: Incremental Memory Injection Jailbreak Attack against Large Language Models

As Large Language Models evolve for user convenience, vulnerability to jailbreak attacks continues to be reported despite ongoing efforts in safety training. Traditional jailbreak techniques typically focus on a single prompt injection, neglecting the models' ability to remember the flow of conversation and the user's instructions. In this paper, we propose Persona Attack, a memory injection based jailbreak method that manipulates the model's context window through a step by…

Read at arXiv (research papers)

arXiv (research papers) Brian Crawford SI

Investigating Detection and Obfuscation of Prompt Injection Attacks Against Software Reverse Engineering AI Agents

Agentic software reverse engineering systems are vulnerable to prompt injection attacks placed into the source code of executable binary files. This research demonstrates defensive tactics for detecting the presences of prompt injection strings in the decompiler output of adversarial example programs. Methods for obfuscating these attacks and subsequent methods for defending against these obfuscations are also explored. This research advances the understanding of risk and…

Read at arXiv (research papers)

Phenomenal World Maria Fernanda Sikorski Geopolitics

Trading Sovereignty

Did the Trump tariffs inadvertently make Lula great again? In 2025, the US unilaterally imposed a 50 percent tax hike on US imports of Brazilian imports in an effort to interfere with the prosecution of Jair Bolsonaro. But rather than capitulation to Trump, what followed was a rapid cycle of Brazilian economic adjustment, with declining exports to the US more than offset through growth in other markets.

Read at Phenomenal World

arXiv (research papers) Nima Dorzhiev SI

Strengthening Polymorphic Prompt Assembling: Dynamic Separator Generation Against Emerging Prompt Injection Attacks

Polymorphic Prompt Assembling (PPA) defends LLM agents against prompt injections by randomly selecting separator pairs from a fixed pool to isolate user input from system instructions. Although effective, static pool reuse exposes a blast-radius vulnerability: once a separator leaks, it can be exploited in future requests. We propose a dynamic per-request separator generation using domain-separated SHA-256 digests keyed on the timestamp, session identifier, and cryptographic…

Read at arXiv (research papers)

arXiv (research papers) Avidan Shah SI

Mitigating Adaptive Attacks against Reasoning Models with Activation Consistency Training

As LLMs gain stronger reasoning capabilities, their extended chain-of-thought introduces new degrees of complexity for defending against adversarial jailbreaks and prompt injection. We study consistency training, a family of fine-tuning objectives that enforce identical behavior on clean prompts and adversarial rewrites, and evaluate its two main variants, output-level (BCT) and activation-level (ACT), across five reasoning models. We formulate both methods as a prompt…

Read at arXiv (research papers)

arXiv (research papers) Ruoqi Guo SI

MIRAGE: Context-Aware Prompt Injection against Mobile GUI Agents via User-Generated Content

Mobile graphical user interface (GUI) agents driven by vision-language models (VLMs) perceive the screen as rendered pixels and choose actions from what they see, so they cannot reliably separate trusted interface elements from user-generated content. We present MIRAGE (Mobile Injection of Realistic Adversarial GUI Examples), a pipeline that turns benign mobile screenshots into prompt-injection samples by placing attacker-controlled text into ordinary user-generated content…

Read at arXiv (research papers)

Journal articles (OpenAlex) Myrna Dawson, Dana Filipetti, Harleen Sangha · Current Sociology Geopolitics

Beijing 30+ and the global movement to prevent femicide and feminicide: The United Nations system as a societal level contributor to social change

The past half-century has been marked by progress and challenges in efforts to curb the global pandemic of male violence against women and girls. This is starkly underscored by the growing global human rights’ movement to reduce sex/gender-related killings, commonly referred to as femicide or feminicide, and recognized as the most brutal and extreme manifestation of such violence. The lives and deaths of women and girls historically and today continue to be structured by…

Read at Journal articles (OpenAlex)

arXiv (research papers) Dongpeng Zhang SI

Localize and Neutralize: Gradient-guided Token Suppression against Visual Prompt Injection Attack

Adversarial images pose a severe security threat to multimodal large language models through prompt injection. Existing defenses largely lack a principled understanding of the underlying mechanisms and struggle to balance efficiency and defense utility. In this work, we show that successful adversarial attacks do not rely on the entire image uniformly but instead depend on a small subset of critical image tokens. Based on this insight, we propose Gradient Token Masking…

Read at arXiv (research papers)

arXiv (research papers) Zixuan Chen SI

IterInject: Indirect Prompt Injection Against LLM Agents via Feedback-Guided Iterative Optimization

LLM-based agents are increasingly deployed for complex tasks requiring planning, tool use, and interaction with external services. Their reliance on untrusted external content exposes them to indirect prompt injection (IPI), in which adversarial instructions embedded in retrieved data hijack agent behavior. Existing attacks rely on static payloads that cannot adapt to agent-specific defenses; even recent adaptive methods lack structured feedback to guide optimization. We…

Read at arXiv (research papers)

arXiv (research papers) Rohan Pandey SI

Poisoning the Watchtower: Prompt Injection Attacks Against LLM-Augmented Security Operations Through Adversarial Log Content

Large language models (LLMs) are increasingly used as analyst assistants in security operations centers (SOCs), where they ingest log and alert data to produce triage labels, incident summaries, or remediation advice. We study a structural failure mode of this design: many log fields are attacker controlled. User agents, URLs, payloads, DNS queries, and attempted usernames can therefore carry instructions to the model alongside evidence of the intrusion. We call this setting…

Read at arXiv (research papers)

arXiv (research papers) Yuanbo Zhou SI

Prompt Overflow: What the Guardrail Inspects Is Not What the Model Infers

Guardrail models (a.k.a. safety checkers) are widely deployed to screen user inputs before they reach large language models (LLMs), serving as a primary defense against prompt injection attacks. Due to strict context constraints, these models handle overlength prompts through truncation or segmentation-based inspection. While prior work has focused on semantic adversarial inputs, the security implications of these long-input processing mechanisms remain largely unexplored…

Read at arXiv (research papers)

Journal articles (OpenAlex) Jonas Gejl Kaas, Rasmus Brun Pedersen, Rasmus Hardt · Cooperation and Conflict Geopolitics

Theorizing small states’ non-alignment strategies under great power rivalry: How Georgia navigates being caught between a rock and a hard place

The article theorizes how neutrality, hedging, and sheltering benefit small states and help them to navigate great power rivalries within buffer complexes. We analyze Georgia’s foreign policy, showing how it evolved from balancing against Russia through Euro–Atlantic alignment to adopting a more flexible combination of the three non-alignment strategies. Contrary to realist expectations and recent small state theorization, Georgia’s case reveals that small states do not…

Read at Journal articles (OpenAlex)

The Guardian Belarus Jakub Krupa

Russian threats against Baltics ‘unacceptable’ and danger to ‘our entire union’, EU’s von der Leyen says - as it happened

This blog is now closed In Brussels, Rutte begins by stressing the importance of Nato ministers meeting in Sweden, the alliance’s “newest ally” who joined in 2024. “This says a great deal about how fundamentally our security environment has changed, especially when it comes to Europe. It is more dangerous, it is more contested and it makes it all the more important that Nato allies work together to safeguard our freedom and security,” he says.

Read at The Guardian Belarus

The Guardian Belarus Jakub Krupa

US warns Russia after Moscow threatens Latvia – as it happened

This blog is now closed, follow the latest development on today’s Europe blog The latest drone alerts come as Ukraine and Latvia were this morning forced by Russia to repeatedly refute Moscow’s claims that Kyiv was preparing attacks against Russia from Latvia. Ukraine’s foreign ministry spokesperson confirmed that Ukraine does not use the territory of Latvia for its operations against Russia and refuted Moscow’s claims. “Russia is lying about Latvia allowing any country to…

Read at The Guardian Belarus

arXiv (research papers) Becky Mashaido SI

On the Geometric Limits of Transformer Defenses against Obfuscation Attacks: Latent Embedding Collapse & Performance Robustness Gap

Prompt injection attacks pose significant risks to language model safety, yet existing defenses are typically evaluated using classification performance. We show that high detection performance does not imply representational robustness. Specifically, multi-operator obfuscated prompts (combining homoglyphs, zero-width characters, and punctuation or emoji noise) can partially collapse onto the embedding manifold of clean prompts, a phenomenon we term latent embedding…

Read at arXiv (research papers)

arXiv (research papers) Lei Zhao SI

LivePI: More Realistic Benchmarking of Agents Against Indirect Prompt Injection

AI agents such as OpenClaw are increasingly deployed in local workflows with access to external tools. This creates indirect prompt-injection (IPI) risk: an agent may execute harmful instructions embedded in untrusted inputs such as email, downloaded files, webpages, repositories, or group-chat messages. Existing evaluations are often small, purely simulated, or focused on a narrow set of channels. We introduce LivePI (Live Prompt Injection), a structured benchmark for IPI…

Read at arXiv (research papers)

Journal articles (OpenAlex) Mathieu Colin, Perrine Hely · Terrorism and Political Violence Geopolitics

“Brothers in Hate”? Hybridization, Convergences and Divisions Between Neo-Nazi Accelerationism and Radical Islam in the Context of the Israeli-Palestinian Conflict

On October 7, 2023, the attack orchestrated by Hamas against Israel and the subsequent wave of reactions within neo-Nazi accelerationist circles provide a privileged vantage point from which to examine the convergences, tensions, and limits of ideological proximity between the extreme right and Islamist movements. Drawing on a qualitative methodology based on the analysis of publications from Telegram channels and accelerationist websites, we propose a tripartite typology of…

Read at Journal articles (OpenAlex)

Older stories