Rosie Zhao
Hi there! I’m a third year PhD student in Computer Science at Harvard University advised by Boaz Barak and Sham Kakade. My research is largely focused on the scientific study of deep learning. I’m grateful to be supported by a Kempner Institute Graduate Fellowship.
I previously completed my undergrad and Master’s at McGill University under the supervision of Prakash Panangaden; there, my research was geared towards representation learning and state abstractions in reinforcement learning, as well as various topics in theoretical computer science.
My socials can be found below.
news
Apr 22, 2025 |
I will be attending ICLR 2025 in Singapore! Come check out our accepted papers on Deconstructing Optimizers for LLM Training and SOAP, or reach out if you want to chat! ![]() |
---|---|
Apr 10, 2025 | New preprint on evaluating the effect of pretraining data on model behavior during RL post-training for mathematical reasoning! See the paper or tweet thread for our ablations across algorithms, hyperparameters, and scale. Done in collaboration with Alexandru Meterez, Sham Kakade, Cengiz Pehlevan, Samy Jelassi, and Eran Malach. 🧠 |
Feb 25, 2025 |
New preprint on studying the nature of emergent capabilities from a distributional lens (across random seeds)! See the paper or tweet thread for our analysis of these bimodal performance distributions on length generalization tasks and natural language settings. Done in collaboration with Tian Qin, David Alvarez-Melis, Sham Kakade, and Naomi Saphra. ![]() |