Hey, I’m Allen Lee. Interpretability researcher and ex-founder (consumer products), based in SF and Taipei. I study what safety training and refusal-editing actually change inside language models.
If you work on making frontier models safer, I’d love to hear from you. I also like talking to people building consumer products with real craft.