I'm a computer science major turned independent interpretability researcher. I've been working to address anomalies and edge cases with sparse autoencoders to improve their performance and usefulness.