Harish Kamath
Hey! I’m Harish. I’m currently on the interpretability team at Anthropic, developing a deeper understanding of how transformers models use attention.
Previously, I worked on 3D computer vision and diffusion models at Scale AI. I graduated from Georgia Tech in 2021, with a Bachelor’s degree in CS & Math.
I believe that AI is the most important technology that humans will ever create. I want to spend my career understanding how it works, and making it useful for everyone.
I love talking to other people excited about their work! Please reach out to me on X if you want to chat.
Other things I like:
- Music (all genres really, both creating and listening)
- Going on walks with friends
- Anything active (I currently like basketball)
Writing
Research
A Microscope into the Dark Matter of Interpretability
We propose a new interpretability architecture that learns a hierarchy of features in LLMs.