Sophie Xhonneux

Leo Schwinn, David Dobre, Sophie Xhonneux et al.

Current research in adversarial robustness of LLMs focuses on discrete input manipulations in the natural language space, which can be directly transferred to closed-source models. However, this approach neglects the steady progression of open-source models. As open-source models advance in capability, ensuring their safety also becomes increasingly imperative. Yet, attacks tailored to open-source LLMs that exploit full model access remain largely unexplored. We address this research gap and propose the embedding space attack, which directly attacks the continuous embedding representation of input tokens. We find that embedding space attacks circumvent model alignments and trigger harmful behaviors more efficiently than discrete attacks or model fine-tuning. Furthermore, we present a novel threat model in the context of unlearning and show that embedding space attacks can extract supposedly deleted information from unlearned LLMs across multiple datasets and models. Our findings highlight embedding space attacks as an important threat model in open-source LLMs. Trigger Warning: the appendix contains LLM-generated text with violence and harassment.

Yan Scholten, Sophie Xhonneux, Leo Schwinn et al.

Current unlearning methods for LLMs optimize on the private information they seek to remove by incorporating it into their fine-tuning data. We argue this not only risks reinforcing exposure to sensitive data, it also fundamentally contradicts the principle of minimizing its use. As a remedy, we propose a novel unlearning method-Partial Model Collapse (PMC), which does not require unlearning targets in the unlearning objective. Our approach is inspired by recent observations that training generative models on their own generations leads to distribution collapse, effectively removing information from model outputs. Our central insight is that model collapse can be leveraged for machine unlearning by deliberately triggering it for data we aim to remove. We theoretically analyze that our approach converges to the desired outcome, i.e. the model unlearns the data targeted for removal. We empirically demonstrate that PMC overcomes three key limitations of existing unlearning methods that explicitly optimize on unlearning targets, and more effectively removes private information from model outputs while preserving general model utility. Overall, our contributions represent an important step toward more comprehensive unlearning that aligns with real-world privacy constraints. Code available at https://www.cs.cit.tum.de/daml/partial-model-collapse/.

Sophie Xhonneux, Alessandro Sordoni, Stephan Günnemann et al.

Large language models (LLMs) are vulnerable to adversarial attacks that can bypass their safety guardrails. In many domains, adversarial training has proven to be one of the most promising methods to reliably improve robustness against such attacks. Yet, in the context of LLMs, current methods for adversarial training are hindered by the high computational costs required to perform discrete adversarial attacks at each training iteration. We address this problem by instead calculating adversarial attacks in the continuous embedding space of the LLM, which is orders of magnitudes more efficient. We propose a fast adversarial training algorithm (C-AdvUL) composed of two losses: the first makes the model robust on continuous embedding attacks computed on an adversarial behaviour dataset; the second ensures the usefulness of the final model by fine-tuning on utility data. Moreover, we introduce C-AdvIPO, an adversarial variant of IPO that does not require utility data for adversarially robust alignment. Our empirical evaluation on five models from different families (Gemma, Phi3, Mistral, Zephyr, Llama2) and at different scales (2B, 3.8B, 7B) shows that both algorithms substantially enhance LLM robustness against discrete attacks (GCG, AutoDAN, PAIR), while maintaining utility. Our results demonstrate that robustness to continuous perturbations can extrapolate to discrete threat models. Thereby, we present a path toward scalable adversarial training algorithms for robustly aligning LLMs.

Tim Beyer, Sophie Xhonneux, Simon Geisler et al.

In this paper, we argue that current safety alignment research efforts for large language models are hindered by many intertwined sources of noise, such as small datasets, methodological inconsistencies, and unreliable evaluation setups. This can, at times, make it impossible to evaluate and compare attacks and defenses fairly, thereby slowing progress. We systematically analyze the LLM safety evaluation pipeline, covering dataset curation, optimization strategies for automated red-teaming, response generation, and response evaluation using LLM judges. At each stage, we identify key issues and highlight their practical impact. We also propose a set of guidelines for reducing noise and bias in evaluations of future attack and defense papers. Lastly, we offer an opposing perspective, highlighting practical reasons for existing limitations. We believe that addressing the outlined problems in future research will improve the field's ability to generate easily comparable results and make measurable progress.

Leo Schwinn, Yan Scholten, Tom Wollschläger et al.

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.

Sophie Xhonneux, David Dobre, Jian Tang et al.

Despite significant investment into safety training, large language models (LLMs) deployed in the real world still suffer from numerous vulnerabilities. One perspective on LLM safety training is that it algorithmically forbids the model from answering toxic or harmful queries. To assess the effectiveness of safety training, in this work, we study forbidden tasks, i.e., tasks the model is designed to refuse to answer. Specifically, we investigate whether in-context learning (ICL) can be used to re-learn forbidden tasks despite the explicit fine-tuning of the model to refuse them. We first examine a toy example of refusing sentiment classification to demonstrate the problem. Then, we use ICL on a model fine-tuned to refuse to summarise made-up news articles. Finally, we investigate whether ICL can undo safety training, which could represent a major security risk. For the safety task, we look at Vicuna-7B, Starling-7B, and Llama2-7B. We show that the attack works out-of-the-box on Starling-7B and Vicuna-7B but fails on Llama2-7B. Finally, we propose an ICL attack that uses the chat template tokens like a prompt injection attack to achieve a better attack success rate on Vicuna-7B and Starling-7B. Trigger Warning: the appendix contains LLM-generated text with violence, suicide, and misinformation.

David Dobre, Mehrnaz Mofakhami, Sophie Xhonneux et al.

Many safety post-training methods for large language models (LLMs) are designed to modify the model's behaviour from producing unsafe answers to issuing refusals. However, such distribution shifts are often brittle and degrade performance on desirable tasks. To address these pitfalls, we propose augmenting the model's vocabulary with a special red flag token, and training the model to insert this token whenever harmful content is generated or imminent. This approach enables the model to explicitly learn the concept of harmfulness in its representations, with minimal impact on utility due to the marginal change in the generated distribution of natural language. Moreover, because the token is embedded in the model's vocabulary, we can naturally leverage the LLMs' generalization capabilities, such as in-context learning (ICL) and out-of-distribution generalization to languages that are not formally supported (e.g., Japanese for Llama3). In particular, we demonstrate that through ICL alone, the model can learn to initiate reflective reasoning upon generating the red flag token at inference, which steers the response away from harmful continuations or enables self-correction when the flag is raised falsely. This approach is orthogonal and complementary to existing safety technique (such as safety classifiers or standard safety training) and easier to evaluate in comparison to natural language refusals, as it does not require a human or automated judge to assess the harmlessness of the answers.