arXiv subjects

Stephen Casper

Publications and source records attributed to Stephen Casper.

At least 37 records · Page 2Linked to original sources

arXiv2026

STACK: Adversarial Attacks on LLM Safeguard Pipelines

Frontier AI developers are relying on layers of safeguards to protect against catastrophic misuse of AI systems. Anthropic and OpenAI guard their latest Opus 4 model and GPT-5 models using such defense pipelines, and other frontier developers including Google DeepMind pledge to soon deploy similar defenses. However, the security of such pipelines is unclear, with limited prior work evaluating or attacking these pipelines. We address this gap by developing and red-teaming an open-source defense pipeline. First, we find that a novel few-shot-prompted input and output classifier outperforms state-of-the-art open-weight safeguard model ShieldGemma across three attacks and two datasets, reducing the attack success rate (ASR) to 0% on the catastrophic misuse dataset ClearHarm. Second, we introduce a STaged AttaCK (STACK) procedure that achieves 71% ASR on ClearHarm in a black-box attack against the few-shot-prompted classifier pipeline. Finally, we also evaluate STACK in a transfer setting, achieving 33% ASR, providing initial evidence that it is feasible to design attacks with no access to the target pipeline. We conclude by suggesting specific mitigations that developers could use to thwart staged attacks.

cs.CL↗

arXiv2026

Expanding External Access To Frontier AI Models For Dangerous Capability Evaluations

Frontier AI companies increasingly rely on external evaluations to assess risks from dangerous capabilities before deployment. However, external evaluators often receive limited model access, limited information, and little time, which can reduce evaluation rigour and confidence. The EU General-Purpose AI Code of Practice calls for "appropriate access", but does not specify what this means in practice. Furthermore, there is no common framework for describing different types and levels of evaluator access. To address this gap, we propose a taxonomy of access methods for dangerous capability evaluations. We disentangle three aspects of access: model access, model information, and evaluation timeframe. For each aspect, we review benefits and risks, including how expanding access can reduce false negatives and improve stakeholder trust, but can also increase security and capacity challenges. We argue that these limitations can likely be mitigated through technical means and safeguards used in other industries. Based on the taxonomy, we propose three descriptive access levels: AL1 (black-box model access and minimal information), AL2 (grey-box model access and substantial information), and AL3 (white-box model access and comprehensive information), to support clearer communication between evaluators, frontier AI companies, and policymakers. We believe these levels correspond to the different standards for appropriate access defined in the EU Code of Practice, though these standards may change over time.

cs.CY↗

arXiv2025

International AI Safety Report 2025: Second Key Update: Technical Safeguards and Risk Management

This second update to the 2025 International AI Safety Report assesses new developments in general-purpose AI risk management over the past year. It examines how researchers, public institutions, and AI developers are approaching risk management for general-purpose AI. In recent months, for example, three leading AI developers applied enhanced safeguards to their new models, as their internal pre-deployment testing could not rule out the possibility that these models could be misused to help create biological weapons. Beyond specific precautionary measures, there have been a range of other advances in techniques for making AI models and systems more reliable and resistant to misuse. These include new approaches in adversarial training, data curation, and monitoring systems. In parallel, institutional frameworks that operationalise and formalise these technical capabilities are starting to emerge: the number of companies publishing Frontier AI Safety Frameworks more than doubled in 2025, and governments and international organisations have established a small number of governance frameworks for general-purpose AI, focusing largely on transparency and risk assessment.

cs.CY↗

arXiv2025

What Features in Prompts Jailbreak LLMs? Investigating the Mechanisms Behind Attacks

Jailbreaks have been a central focus of research regarding the safety and reliability of large language models (LLMs), yet the mechanisms underlying these attacks remain poorly understood. While previous studies have predominantly relied on linear methods to detect jailbreak attempts and model refusals, we take a different approach by examining both linear and non-linear features in prompts that lead to successful jailbreaks. First, we introduce a novel dataset comprising 10,800 jailbreak attempts spanning 35 diverse attack methods. Leveraging this dataset, we train linear and non-linear probes on hidden states of open-weight LLMs to predict jailbreak success. Probes achieve strong in-distribution accuracy but transfer is attack-family-specific, revealing that different jailbreaks are supported by distinct internal mechanisms rather than a single universal direction. To establish causal relevance, we construct probe-guided latent interventions that systematically shift compliance in the predicted direction. Interventions derived from non-linear probes produce larger and more reliable effects than those from linear probes, indicating that features linked to jailbreak success are encoded non-linearly in prompt representations. Overall, the results surface heterogeneous, non-linear structure in jailbreak mechanisms and provide a prompt-side methodology for recovering and testing the features that drive jailbreak outcomes.

cs.CR↗

arXiv2025

Pitfalls of Evidence-Based AI Policy

Nations across the world are working to govern AI. However, from a technical perspective, there is uncertainty and disagreement on the best way to do this. Meanwhile, recent debates over AI regulation have led to calls for "evidence-based AI policy" which emphasize holding regulatory action to a high evidentiary standard. Evidence is of irreplaceable value to policymaking. However, holding regulatory action to too high an evidentiary standard can lead to systematic neglect of certain risks. In historical policy debates (e.g., over tobacco ca. 1965 and fossil fuels ca. 1985) "evidence-based policy" rhetoric is also a well-precedented strategy to downplay the urgency of action, delay regulation, and protect industry interests. Here, we argue that if the goal is evidence-based AI policy, the first regulatory objective must be to actively facilitate the process of identifying, studying, and deliberating about AI risks. We discuss a set of 15 regulatory goals to facilitate this and show that Brazil, Canada, China, the EU, South Korea, the UK, and the USA all have substantial opportunities to adopt further evidence-seeking policies.

cs.CY↗

arXiv2025

Audit Cards: Contextualizing AI Evaluations

AI governance frameworks increasingly rely on audits, yet the results of their underlying evaluations require interpretation and context to be meaningfully informative. Even technically rigorous evaluations can offer little useful insight if reported selectively or obscurely. Current literature focuses primarily on technical best practices, but evaluations are an inherently sociotechnical process, and there is little guidance on reporting procedures and context. Through literature review, stakeholder interviews, and analysis of governance frameworks, we propose "audit cards" to make this context explicit. We identify six key types of contextual features to report and justify in audit cards: auditor identity, evaluation scope, methodology, resource access, process integrity, and review mechanisms. Through analysis of existing evaluation reports, we find significant variation in reporting practices, with most reports omitting crucial contextual information such as auditors' backgrounds, conflicts of interest, and the level and type of access to models. We also find that most existing regulations and frameworks lack guidance on rigorous reporting. In response to these shortcomings, we argue that audit cards can provide a structured format for reporting key claims alongside their justifications, enhancing transparency, facilitating proper interpretation, and establishing trust in reporting.

cs.CY↗

arXiv2025

Defending Against Unforeseen Failure Modes with Latent Adversarial Training

Despite extensive diagnostics and debugging by developers, AI systems sometimes exhibit harmful unintended behaviors. Finding and fixing these is challenging because the attack surface is so large -- it is not tractable to exhaustively search for inputs that may elicit harmful behaviors. Red-teaming and adversarial training (AT) are commonly used to improve robustness, however, they empirically struggle to fix failure modes that differ from the attacks used during training. In this work, we utilize latent adversarial training (LAT) to defend against vulnerabilities without leveraging knowledge of what they are or using inputs that elicit them. LAT makes use of the compressed, abstract, and structured latent representations of concepts that the network actually uses for prediction. Here, we use it to defend against failure modes without examples that elicit them. Specifically, we use LAT to remove backdoors and defend against held-out classes of adversarial attacks. We show in image classification, text classification, and text generation tasks that LAT usually improves both robustness to novel attacks and performance on clean data relative to AT. This suggests that LAT can be a promising tool for defending against failure modes that are not explicitly identified by developers.

cs.CR↗

arXiv2025

Latent Adversarial Training Improves Robustness to Persistent Harmful Behaviors in LLMs

Large language models (LLMs) can often be made to behave in undesirable ways that they are explicitly fine-tuned not to. For example, the LLM red-teaming literature has produced a wide variety of 'jailbreaking' techniques to elicit harmful text from models that were fine-tuned to be harmless. Recent work on red-teaming, model editing, and interpretability suggests that this challenge stems from how (adversarial) fine-tuning largely serves to suppress rather than remove undesirable capabilities from LLMs. Prior work has introduced latent adversarial training (LAT) as a way to improve robustness to broad classes of failures. These prior works have considered untargeted latent space attacks where the adversary perturbs latent activations to maximize loss on examples of desirable behavior. Untargeted LAT can provide a generic type of robustness but does not leverage information about specific failure modes. Here, we experiment with targeted LAT where the adversary seeks to minimize loss on a specific competing task. We find that it can augment a wide variety of state-of-the-art methods. First, we use targeted LAT to improve robustness to jailbreaks, outperforming a strong R2D2 baseline with orders of magnitude less compute. Second, we use it to more effectively remove backdoors with no knowledge of the trigger. Finally, we use it to more effectively unlearn knowledge for specific undesirable tasks in a way that is also more robust to re-learning. Overall, our results suggest that targeted LAT can be an effective tool for defending against harmful behaviors from LLMs.

cs.LG↗

arXiv2025

Model Tampering Attacks Enable More Rigorous Evaluations of LLM Capabilities

Evaluations of large language model (LLM) risks and capabilities are increasingly being incorporated into AI risk management and governance frameworks. Currently, most risk evaluations are conducted by designing inputs that elicit harmful behaviors from the system. However, this approach suffers from two limitations. First, input-output evaluations cannot fully evaluate realistic risks from open-weight models. Second, the behaviors identified during any particular input-output evaluation can only lower-bound the model's worst-possible-case input-output behavior. As a complementary method for eliciting harmful behaviors, we propose evaluating LLMs with model tampering attacks which allow for modifications to latent activations or weights. We pit state-of-the-art techniques for removing harmful LLM capabilities against a suite of 5 input-space and 6 model tampering attacks. In addition to benchmarking these methods against each other, we show that (1) model resilience to capability elicitation attacks lies on a low-dimensional robustness subspace; (2) the success rate of model tampering attacks can empirically predict and offer conservative estimates for the success of held-out input-space attacks; and (3) state-of-the-art unlearning methods can easily be undone within 16 steps of fine-tuning. Together, these results highlight the difficulty of suppressing harmful LLM capabilities and show that model tampering attacks enable substantially more rigorous evaluations than input-space attacks alone.

cs.CR↗

arXiv2025

The Singapore Consensus on Global AI Safety Research Priorities

Rapidly improving AI capabilities and autonomy hold significant promise of transformation, but are also driving vigorous debate on how to ensure that AI is safe, i.e., trustworthy, reliable, and secure. Building a trusted ecosystem is therefore essential -- it helps people embrace AI with confidence and gives maximal space for innovation while avoiding backlash. The "2025 Singapore Conference on AI (SCAI): International Scientific Exchange on AI Safety" aimed to support research in this space by bringing together AI scientists across geographies to identify and synthesise research priorities in AI safety. This resulting report builds on the International AI Safety Report chaired by Yoshua Bengio and backed by 33 governments. By adopting a defence-in-depth model, this report organises AI safety research domains into three types: challenges with creating trustworthy AI systems (Development), challenges with evaluating their risks (Assessment), and challenges with monitoring and intervening after deployment (Control).

cs.AI↗

arXiv2025

Open Problems in Technical AI Governance

AI progress is creating a growing range of risks and opportunities, but it is often unclear how they should be navigated. In many cases, the barriers and uncertainties faced are at least partly technical. Technical AI governance, referring to technical analysis and tools for supporting the effective governance of AI, seeks to address such challenges. It can help to (a) identify areas where intervention is needed, (b) identify and assess the efficacy of potential governance actions, and (c) enhance governance options by designing mechanisms for enforcement, incentivization, or compliance. In this paper, we explain what technical AI governance is, why it is important, and present a taxonomy and incomplete catalog of its open problems. This paper is intended as a resource for technical researchers or research funders looking to contribute to AI governance.

cs.CY↗

arXiv2025

International Scientific Report on the Safety of Advanced AI (Interim Report)

This is the interim publication of the first International Scientific Report on the Safety of Advanced AI. The report synthesises the scientific understanding of general-purpose AI -- AI that can perform a wide variety of tasks -- with a focus on understanding and managing its risks. A diverse group of 75 AI experts contributed to this report, including an international Expert Advisory Panel nominated by 30 countries, the EU, and the UN. Led by the Chair, these independent experts collectively had full discretion over the report's content. The final report is available at arXiv:2501.17805

cs.CY↗

arXiv2025

Randomness, Not Representation: The Unreliability of Evaluating Cultural Alignment in LLMs

Research on the 'cultural alignment' of Large Language Models (LLMs) has emerged in response to growing interest in understanding representation across diverse stakeholders. Current approaches to evaluating cultural alignment through survey-based assessments that borrow from social science methodologies often overlook systematic robustness checks. Here, we identify and test three assumptions behind current survey-based evaluation methods: (1) Stability: that cultural alignment is a property of LLMs rather than an artifact of evaluation design, (2) Extrapolability: that alignment with one culture on a narrow set of issues predicts alignment with that culture on others, and (3) Steerability: that LLMs can be reliably prompted to represent specific cultural perspectives. Through experiments examining both explicit and implicit preferences of leading LLMs, we find a high level of instability across presentation formats, incoherence between evaluated versus held-out cultural dimensions, and erratic behavior under prompt steering. We show that these inconsistencies can cause the results of an evaluation to be very sensitive to minor variations in methodology. Finally, we demonstrate in a case study on evaluation design that narrow experiments and a selective assessment of evidence can be used to paint an incomplete picture of LLMs' cultural alignment properties. Overall, these results highlight significant limitations of current survey-based approaches to evaluating the cultural alignment of LLMs and highlight a need for systematic robustness checks and red-teaming for evaluation results. Data and code are available at https://huggingface.co/datasets/akhan02/cultural-dimension-cover-letters and https://github.com/ariba-k/llm-cultural-alignment-evaluation, respectively.

cs.CY↗

arXiv2025

Adversarial Alignment for LLMs Requires Simpler, Reproducible, and More Measurable Objectives

Misaligned research objectives have considerably hindered progress in adversarial robustness research over the past decade. For instance, an extensive focus on optimizing target metrics, while neglecting rigorous standardized evaluation, has led researchers to pursue ad-hoc heuristic defenses that were seemingly effective. Yet, most of these were exposed as flawed by subsequent evaluations, ultimately contributing little measurable progress to the field. In this position paper, we illustrate that current research on the robustness of large language models (LLMs) risks repeating past patterns with potentially worsened real-world implications. To address this, we argue that realigned objectives are necessary for meaningful progress in adversarial alignment. To this end, we build on established cybersecurity taxonomy to formally define differences between past and emerging threat models that apply to LLMs. Using this framework, we illustrate that progress requires disentangling adversarial alignment into addressable sub-problems and returning to core academic principles, such as measureability, reproducibility, and comparability. Although the field presents significant challenges, the fresh start on adversarial robustness offers the unique opportunity to build on past experience while avoiding previous mistakes.

cs.LG↗

arXiv2025

Obfuscated Activations Bypass LLM Latent-Space Defenses

Recent latent-space monitoring techniques have shown promise as defenses against LLM attacks. These defenses act as scanners that seek to detect harmful activations before they lead to undesirable actions. This prompts the question: Can models execute harmful behavior via inconspicuous latent states? Here, we study such obfuscated activations. We show that state-of-the-art latent-space defenses -- including sparse autoencoders, representation probing, and latent OOD detection -- are all vulnerable to obfuscated activations. For example, against probes trained to classify harmfulness, our attacks can often reduce recall from 100% to 0% while retaining a 90% jailbreaking rate. However, obfuscation has limits: we find that on a complex task (writing SQL code), obfuscation reduces model performance. Together, our results demonstrate that neural activations are highly malleable: we can reshape activation patterns in a variety of ways, often while preserving a network's behavior. This poses a fundamental challenge to latent-space defenses.

cs.LG↗

arXiv2025

The AI Agent Index

Leading AI developers and startups are increasingly deploying agentic AI systems that can plan and execute complex tasks with limited human involvement. However, there is currently no structured framework for documenting the technical components, intended uses, and safety features of agentic systems. To fill this gap, we introduce the AI Agent Index, the first public database to document information about currently deployed agentic AI systems. For each system that meets the criteria for inclusion in the index, we document the system's components (e.g., base model, reasoning implementation, tool use), application domains (e.g., computer use, software engineering), and risk management practices (e.g., evaluation results, guardrails), based on publicly available information and correspondence with developers. We find that while developers generally provide ample information regarding the capabilities and applications of agentic systems, they currently provide limited information regarding safety and risk management practices. The AI Agent Index is available online at https://aiagentindex.mit.edu/

cs.SE↗

arXiv2025

International AI Safety Report

The first International AI Safety Report comprehensively synthesizes the current evidence on the capabilities, risks, and safety of advanced AI systems. The report was mandated by the nations attending the AI Safety Summit in Bletchley, UK. Thirty nations, the UN, the OECD, and the EU each nominated a representative to the report's Expert Advisory Panel. A total of 100 AI experts contributed, representing diverse perspectives and disciplines. Led by the report's Chair, these independent experts collectively had full discretion over the report's content.

cs.CY↗

arXiv2025

Open Problems in Mechanistic Interpretability

Mechanistic interpretability aims to understand the computational mechanisms underlying neural networks' capabilities in order to accomplish concrete scientific and engineering goals. Progress in this field thus promises to provide greater assurance over AI system behavior and shed light on exciting scientific questions about the nature of intelligence. Despite recent progress toward these goals, there are many open problems in the field that require solutions before many scientific and practical benefits can be realized: Our methods require both conceptual and practical improvements to reveal deeper insights; we must figure out how best to apply our methods in pursuit of specific goals; and the field must grapple with socio-technical challenges that influence and are influenced by our work. This forward-facing review discusses the current frontier of mechanistic interpretability and the open problems that the field may benefit from prioritizing.

cs.LG↗