My recent work involves improving the capabilities of large language models and better aligning them to human values. Here are some of my representative publications (see all).
Long-form factuality in large language models. [code]
Symbol tuning improves in-context learning in language models. [code]
Simple synthetic data reduces sycophancy in large language models. [code]
Larger language models do in-context learning differently. [blog]
Generative image translation for data augmentation in colorectal histopathology images. [talk]