Kenneth Li

About

Kenneth Li is a third-year PhD student at Harvard University. He is advised by Fernanda Viégas, Hanspeter Pfister, and Martin Wattenberg and is supported by the Kempner Institute Graduate Fellowship. Li aims at understanding the inner workings of large language models and, based on these findings, improving the controllability of current model behavior to secure human benefits.

Research Focus

Li’s current research focuses on ML, NLP, and Interpretability. He aims at understanding the inner workings of large language models and, based on these findings, improving the controllability of current model behavior to secure human benefit. “Kempner provides great support for my further exploration of the internal workings of large language models,” Li says. “While it evokes conventional neuroscience and psychology, we entertain full knowledge and control of what every neuron and synapse is doing and we can run experiments economically and free of ethical concerns.” Based on the understanding that Li has gained from interpretability research, he wishes to devise techniques that could potentially align language models with human interests.