Aaron Mueller

Priors in time: Missing inductive biases for language model interpretability

Discovering Latent Biases in Language Models with Steering Vectors

From Isolation to Entanglement: When Do Interpretability Methods Identify and Disentangle Known Concepts?

In-Context Learning Without Copying

Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models

Arithmetic Without Algorithms: Language Models Solve Math with a Bag of Heuristics

NNsight and NDIF: Democratizing Access to Open-Weight Foundation Model Internals

MIB: A Mechanistic Interpretability Benchmark