2025
Efficient Automated Circuit Discovery in Transformers using Contextual Decomposition
Aliyah R. Hsu, Georgia Zhou, Yeshwanth Cherapanamjeri, Yaxuan Huang, Anobel Odisho, Peter R. Carroll +1
ICLR 2025poster
Automated mechanistic interpretation research has attracted great interest due to its potential to scale explanations of neural network internals to large models. Existing automated circuit discovery work relies on activation patching or its approximations to identify subgraphs in models for specifi…