Neuron to Graph
Interpreting Language Model Neurons at Scale
- Alex Foote1*, Neel Nanda2, Esben Kran1, Ionnis Konstas3, Shay Cohen4, Fazl Barez1,4,5*
1Apart Research 2Independent 3School of Mathematical and Computer Sciences Heriot-Watt University 4School of Informatics, University of Edinburgh 5University of Oxford
* Equal contribution
Abstract
Understanding the function of individual neurons within language models is essential for mechanistic interpretability research. We propose Neuron to Graph (N2G), a tool which takes a neuron and its dataset examples, and automatically distills the neuron's behaviour on those examples to an interpretable graph. This presents a less labour intensive approach to interpreting neurons than current manual methods, that will better scale these methods to large language models (LLMs). We use truncation and saliency methods to only present the important tokens, and augment the dataset examples with more diverse samples to better capture the extent of neuron behaviour. These graphs can be visualised to aid manual interpretation by researchers, but can also output token activations on text to compare to the neuron's ground truth activations for automatic validation. N2G represents a step towards scalable interpretability methods by allowing us to convert neurons in an LLM to interpretable representations of measurable quality.
Work in progress