相关论文: Steered Generation via Gradient-Based Optimization…

Recent LLMs like DeepSeek-R1 have demonstrated state-of-the-art performance by integrating deep thinking and complex reasoning during generation. However, the internal mechanisms behind these reasoning processes remain unexplored. We…

Large language models possess strong chemical reasoning capabilities, making them effective molecular editors. However, property-relevant information is implicitly entangled across their dense hidden states, providing no explicit handle for…

Steering has emerged as a promising approach in controlling large language models (LLMs) without modifying model parameters. However, most existing steering methods rely on large-scale datasets to learn clear behavioral information, which…

Sparse Autoencoders (SAEs) have been proposed as an unsupervised approach to learn a decomposition of a model's latent space. This enables useful applications such as steering - influencing the output of a model towards a desired concept -…

Sparse Autoencoders uncover thousands of features in vision models, yet explaining these features without requiring human intervention remains an open challenge. While previous work has proposed generating correlation-based explanations…

We propose a novel method that leverages sparse autoencoders (SAEs) and clustering techniques to analyze the internal token representations of large language models (LLMs) and guide generations in mathematical reasoning tasks. Our approach…

Sparsity is a desirable attribute. It can lead to more efficient and more effective representations compared to the dense model. Meanwhile, learning sparse latent representations has been a challenging problem in the field of computer…

Recent developments in Large Language Model (LLM) capabilities have brought great potential but also posed new risks. For example, LLMs with knowledge of bioweapons, advanced chemistry, or cyberattacks could cause violence if placed in the…

Recent work in Mechanistic Interpretability (MI) has enabled the identification and intervention of internal features in Large Language Models (LLMs). However, a persistent challenge lies in linking such internal features to the reliable…

Large Language Models (LLMs) encode factual knowledge within hidden parametric spaces that are difficult to inspect or control. While Sparse Autoencoders (SAEs) can decompose hidden activations into more fine-grained, interpretable…

Sparse Autoencoders (SAEs) are widely employed for mechanistic interpretability and model steering. Within this context, steering is by design performed by means of decoding altered SAE intermediate representations. This procedure…

Sparse autoencoders (SAEs) are a popular method for interpreting concepts represented in large language model (LLM) activations. However, there is a lack of evidence regarding the validity of their interpretations due to the lack of a…

Dynamic head pruning in Vision Transformers (ViTs) improves efficiency by removing redundant attention heads, but existing pruning policies are often difficult to interpret and control. In this work, we propose a novel framework by…

Sparse autoencoders (SAEs) are a useful tool for uncovering human-interpretable features in the activations of large language models (LLMs). While some expect SAEs to find the true underlying features used by a model, our research shows…

Sparse autoencoders (SAEs) provide a powerful mechanism for decomposing the dense representations produced by Large Language Models (LLMs) into interpretable latent features. We posit that SAEs constitute a natural foundation for Learned…

Internal activations of diffusion models encode rich semantic information, but interpreting such representations remains challenging. While Sparse Autoencoders (SAEs) have shown promise in disentangling latent representations, existing…

Graph-based surrogate models provide fast alternatives to high-fidelity CFD solvers, but their opaque latent spaces and limited controllability restrict use in safety-critical settings. A key failure mode in oscillatory flows is phase…

We study how reliably sparse autoencoders (SAEs) support claims about reasoning-related internal features in large language models. We first give a stylized analysis showing that sparsity-regularized decoding can preferentially retain…

Responsible deployment of language models requires mechanisms for refusing unsafe prompts while preserving model performance. While most approaches modify model weights through additional training, we explore an alternative: steering model…

Understanding the internal representations of large language models (LLMs) remains a central challenge for interpretability research. Sparse autoencoders (SAEs) offer a promising solution by decomposing activations into interpretable…