Independent Mechanistic Interpretability Research with supervision from Neel Nanda and Swarat Chaudhuri. Published paper at ICML 2024 on self-repair.

- First Authored Copy Suppression Mechanistic Interpretability paper (in review at ICLR) explaining 76.9% of an Attention Head in a Large Language Model as well as its connection to Self-Repair- Full-time research mentorship from Neel Nanda (Deepmind) on mechanistic interpretability; used PyTorch/TransformerLens to uncover novel phenomena in Transformers- Reverse-engineered preliminary circuitry for the completion of ‘dual pairs’ of words in GPT-2 Small.

- Imaging and setup of new computers for new employee arrivals- Training and utilization of Burp Suite, Gobuster, and Nmap penetration testing software (among many)- Development of new employee equipment imaging playbooks- Vulnerability scanning of equipment for compliance to CIS standards- Completion of Google's "The Bits and Bytes of Computer Networking" course- Configuration and daily administration of Basecamp project management software- Presentation of project management software comparison and features- Configuration, development and daily administration of 15 IT/Security projects in SmartsheetsI am very thankful for the opportunity to attend this fantastic experience, and it was a pleasure working with everyone!