arXiv subjects

Lisa Thiergart

Publications and source records attributed to Lisa Thiergart.

5 recordsLinked to original sources

arXiv2026

SL5 Standard for AI Security

Security Level 5 (SL5) is a security posture for AI systems that could plausibly thwart top-priority operations by the world's most cyber-capable institutions: those with extensive resources, state-level infrastructure, and expertise years ahead of the public state of the art. The SL5 terminology originates from the RAND Corporation's 2024 report "Securing AI Model Weights". Frontier AI development requires use-case-specific, productivity-optimised and updateable AI datacenter security standards. This first revision of the SL5 standard focuses on requirements with long lead times: interventions that must be planned years in advance, such as facility construction, hardware procurement, and organizational capability development. We prioritize these requirements because preserving optionality for SL5 by 2028/2029 requires starting now. These capabilities cannot be retrofitted on short notice when the need becomes urgent. Some requirements represent significant departures from current day standard practice. We believe bold measures are necessary for this level of security and see clear opportunities to apply optimization pressure to existing and novel solutions to customize them for the AI industry and address the practical operational requirements as much as possible. Our organization exists to begin paving this path. Some requirements approximate government security capabilities where private-sector approaches may be insufficient. We identify these gaps and note where government involvement may ultimately be necessary.

cs.CR↗

arXiv2025

Mechanisms to Verify International Agreements About AI Development

International agreements about AI development may be required to reduce catastrophic risks from advanced AI systems. However, agreements about such a high-stakes technology must be backed by verification mechanisms--processes or tools that give one party greater confidence that another is following the agreed-upon rules, typically by detecting violations. This report gives an overview of potential verification approaches for three example policy goals, aiming to demonstrate how countries could practically verify claims about each other's AI development and deployment. The focus is on international agreements and state-involved AI development, but these approaches could also be applied to domestic regulation of companies. While many of the ideal solutions for verification are not yet technologically feasible, we emphasize that increased access (e.g., physical inspections of data centers) can often substitute for these technical approaches. Therefore, we remain hopeful that significant political will could enable ambitious international coordination, with strong verification mechanisms, to reduce catastrophic AI risks.

cs.CY↗

arXiv2024

What AI evaluations for preventing catastrophic risks can and cannot do

AI evaluations are an important component of the AI governance toolkit, underlying current approaches to safety cases for preventing catastrophic risks. Our paper examines what these evaluations can and cannot tell us. Evaluations can establish lower bounds on AI capabilities and assess certain misuse risks given sufficient effort from evaluators. Unfortunately, evaluations face fundamental limitations that cannot be overcome within the current paradigm. These include an inability to establish upper bounds on capabilities, reliably forecast future model capabilities, or robustly assess risks from autonomous AI systems. This means that while evaluations are valuable tools, we should not rely on them as our main way of ensuring AI systems are safe. We conclude with recommendations for incremental improvements to frontier AI safety, while acknowledging these fundamental limitations remain unsolved.

cs.CY↗

arXiv2024

Declare and Justify: Explicit assumptions in AI evaluations are necessary for effective regulation

As AI systems advance, AI evaluations are becoming an important pillar of regulations for ensuring safety. We argue that such regulation should require developers to explicitly identify and justify key underlying assumptions about evaluations as part of their case for safety. We identify core assumptions in AI evaluations (both for evaluating existing models and forecasting future models), such as comprehensive threat modeling, proxy task validity, and adequate capability elicitation. Many of these assumptions cannot currently be well justified. If regulation is to be based on evaluations, it should require that AI development be halted if evaluations demonstrate unacceptable danger or if these assumptions are inadequately justified. Our presented approach aims to enhance transparency in AI development, offering a practical path towards more effective governance of advanced AI systems.

cs.AI↗

arXiv2023

Steering Language Models With Activation Engineering

Prompt engineering and finetuning aim to maximize language model performance on a given metric (like toxicity reduction). However, these methods do not fully elicit a model's capabilities. To reduce this gap, we introduce activation engineering: the inference-time modification of activations in order to control (or steer) model outputs. Specifically, we introduce the Activation Addition (ActAdd) technique, which contrasts the intermediate activations on prompt pairs (such as "Love" versus "Hate") to compute a steering vector (Subramani et al. 2022). By tactically adding in e.g. the "Love" - "Hate" steering vector during the forward pass, we achieve SOTA on negative-to-positive sentiment shift and detoxification using models including LLaMA-3 and OPT. ActAdd yields inference-time control over high-level output properties (like topic and sentiment) while preserving performance on off-target tasks. ActAdd is lightweight: it does not require any machine optimization and works with a single pair of data points, which enables rapid iteration over steering. ActAdd demonstrates the power of activation engineering.

cs.CL↗