BrainInsideTheMachine
Mechanistic interpretability of multilingual reasoning in transformers. 170+ causal intervention experiments across 4 model families.
- 1
- Jupyter Notebook
- 4 mo ago
GregSharma
M.S. in Mathematics in Finance @ NYU Courant
Built from GregSharma's public GitHub profile, New York, New York. Not on Take My Repo yet, so there are no posts or updates here.
Mechanistic interpretability of multilingual reasoning in transformers. 170+ causal intervention experiments across 4 model families.
Ready-to-run Google Colab notebooks for compute-flexible jobs