Studies show that safety-tuned models may nevertheless divulge harmful information. In this
work, we show that whether they do so depends significantly on who they are talking to …

Sparse autoencoders (SAEs) have become a core ingredient in the reverse engineering of
large-language models (LLMs). For LLMs, they have been shown to decompose …

Mechanistic interpretability aims to understand the computational mechanisms underlying
neural networks' capabilities in order to accomplish concrete scientific and engineering …

A central question in multilingual language modeling is whether large language models
(LLMs) develop a universal concept representation, disentangled from specific languages …

We propose SelfControl, an inference-time model control method utilizing gradients to
control the behavior of large language models (LLMs) without explicit human annotations …

Interpretability methods seek to understand language model representations, yet the outputs
of most such methods--circuits, vectors, scalars--are not immediately human-interpretable. In …

Understanding the internal computations of large language models (LLMs) is crucial for
aligning them with human values and preventing undesirable behaviors like toxic content …

A central question in multilingual language modeling is whether large language models
(LLMs) develop a universal concept representation, disentangled from specific languages …

When making predictions, a language model must trade off how much it relies on its context
vs. its prior knowledge. Choosing how sensitive the model is to its context is a fundamental …

This study presents a novel approach that leverages Neural Ordinary Differential Equations
(Neural ODEs) to unravel the intricate relationships between inputs and outputs in Large …