8 papers
Manifold Steering Reveals the Shared Geometry of Neural Network Representation and Behavior
Daniel Wurgaft, Can Rager, Matthew Kowal +13
Neural representations carry rich geometric structure; but does that structure causally shape behavior? To address this question, we intervene along paths through activation space…
Arithmetic in the Wild: Llama uses Base-10 Addition to Reason About Cyclic Concepts
Sheridan Feucht, Tal Haklay, Usha Bhalla +9
Does structure in representations imply structure in computation? We study how Llama-3.1-8B reasons over cyclic concepts (e.g., "what month is six months after August?"). Even thou…
Do Sparse Autoencoders Capture Concept Manifolds?
Usha Bhalla, Thomas Fel, Can Rager +9
Sparse autoencoders (SAEs) are widely used to extract interpretable features from neural network representations, often under the implicit assumption that concepts correspond to in…
In-Context Learning Without Copying
Kerem Sahin, Sheridan Feucht, Adam Belfki +4
Induction heads are attention heads that perform inductive copying by matching patterns from earlier context and copying their continuations verbatim. As models develop induction h…
Vector Arithmetic in Concept and Token Subspaces
Sheridan Feucht, Byron Wallace, David Bau
In order to predict the next token, LLMs must represent semantic and surface-level information about the current word. Previous work identified two types of attention heads that di…
Deep Neural Networks Can Learn Generalizable Same-Different Visual Relations
Alexa R. Tartaglini, Sheridan Feucht, Michael A. Lepori +4
Although deep neural networks can achieve human-level performance on many object recognition benchmarks, prior work suggests that these same models fail to learn simple abstract re…