Steering methods for language models (LMs) seek to provide fine-grained and interpretable control over model generations by variously changing model inputs, weights, or representations to adjust behavior. Recent work has shown that adjusting weights or representations is often less effective than steering by prompting, for instance when wanting to introduce or suppress a particular concept. We demonstrate how to improve representation steering via our new Reference-free Preference Steering (RePS), a bidirectional preference-optimization objective that jointly does concept steering and suppression. We train three parameterizations of RePS and evaluate them on AXBENCH, a large-scale model steering benchmark. On Gemmamodels with sizes ranging from 2Bto 27B, RePS outperforms all existing steering methods trained with a language modeling objective and substantially narrows the gap with prompting - while promoting interpretability and minimizing parameter count. In suppression, RePS matches the language-modeling objective on Gemma-2 and outperforms it on the larger Gemma-3 variants while remaining resilient to prompt-based jailbreaking attacks that defeat prompting. Overall, our results suggest that RePS provides an interpretable and robust alternative to prompting for both steering and suppression.

We propose a novel spectral generative modeling framework for natural language processing that jointly learns a global time varying Fourier dictionary and per token mixing coefficients, replacing the ubiquitous self attention mechanism in transformer architectures. By enforcing reconstruction losses in both the time domain (embedding reconstruction) and the frequency domain (via Short Time Fourier Transform magnitude matching) alongside a standard language modeling objective, and fitting a Gaussian Mixture Model (GMM) prior over the learned mixing vectors, our approach achieves competitive perplexity and generation quality on standard benchmarks such as WikiText2 and Penn Treebank. In contrast to the quadratic computation complexity of self attention, our method operates with linear complexity, delivering substantial efficiency gains. We demonstrate that spectral dictionary models can achieve competitive performance compared to transformer baselines while significantly reducing inference latency and memory footprint, offering a compelling alternative for scalable language modeling.

Large Language Models (LLMs) are swiftly advancing in architecture and capability, and as they integrate more deeply into complex systems, the urgency to scrutinize their security properties grows. This paper surveys research in the emerging interdisciplinary field of adversarial attacks on LLMs, a subfield of trustworthy ML, combining the perspectives of Natural Language Processing and Security. Prior work has shown that even safety-aligned LLMs (via instruction tuning and reinforcement learning through human feedback) can be susceptible to adversarial attacks, which exploit weaknesses and mislead AI systems, as evidenced by the prevalence of `jailbreak' attacks on models like ChatGPT and Bard. In this survey, we first provide an overview of large language models, describe their safety alignment, and categorize existing research based on various learning structures: textual-only attacks, multi-modal attacks, and additional attack methods specifically targeting complex systems, such as federated learning or multi-agent systems. We also offer comprehensive remarks on works that focus on the fundamental sources of vulnerabilities and potential defenses. To make this field more accessible to newcomers, we present a systematic review of existing works, a structured typology of adversarial attack concepts, and additional resources, including slides for presentations on related topics at the 62nd Annual Meeting of the Association for Computational Linguistics (ACL'24).

While the successes of transformers across many domains are indisputable, accurate understanding of the learning mechanics is still largely lacking. Their capabilities have been probed on benchmarks which include a variety of structured and reasoning tasks -- but mathematical understanding is lagging substantially behind. Recent lines of work have begun studying representational aspects of this question: that is, the size/depth/complexity of attention-based networks to perform certain tasks. However, there is no guarantee the learning dynamics will converge to the constructions proposed. In our paper, we provide fine-grained mechanistic understanding of how transformers learn "semantic structure", understood as capturing co-occurrence structure of words. Precisely, we show, through a combination of mathematical analysis and experiments on Wikipedia data and synthetic data modeled by Latent Dirichlet Allocation (LDA), that the embedding layer and the self-attention layer encode the topical structure. In the former case, this manifests as higher average inner product of embeddings between same-topic words. In the latter, it manifests as higher average pairwise attention between same-topic words. The mathematical results involve several assumptions to make the analysis tractable, which we verify on data, and might be of independent interest as well.

In recent years, language models (LMs) have become more prominent in natural language processing (NLP) research and are also becoming increasingly impactful in practice. Scaling up LMs has been shown to improve performance across a range of NLP tasks. For instance, scaling up language models can improve perplexity across seven orders of magnitude of model sizes, and new abilities such as multi-step reasoning have been observed to arise as a result of model scale. However, one of the challenges of continued scaling is that training new, larger models requires great amounts of computational resources. Moreover, new models are often trained from scratch and do not leverage the weights from previously existing models.

In our study, we show how a Transformer language model, trained simply to predict a masked (hidden) amino acid in a protein sequence, recovers high-level structural and functional properties of proteins through its attention mechanism. We demonstrate that attention (1) captures the folding structure of proteins, connecting regions that are apart in the underlying sequence but spatially close in the protein structure, and (2) targets binding sites, a key functional component of proteins. We also introduce a three-dimensional visualization of the interaction between attention and protein structure. Our findings align with biological processes and provide a tool to aid scientific discovery. Proteins are complex molecules that play a critical functional and structural role for all forms of life on this planet. The study of proteins has led to many advances in disease therapies, and the application of machine learning to proteins has the potential for far-reaching applications in medicine and beyond.