Pivotal Token Search

Adversarial Manipulation of CoT

Analysed determinism, faithfulness, reasoning patterns, & steering. Developed and tested methods to enhance control and fail-safes

Mechanistic analysis of Chain-of-Thought faithfulness using GPT-2 Small

Implementation and analysis of Sparse Autoencoders for neural network interpretability research. Features interactive visualization dashboard and W&B integration.

My AI interpretability research journey

Unofficial implementation to reproduce the experiments from "Superposition as a Phase Change" of "Toy Models of Superposition".