Jacob Springer
[cv] [google scholar] [github] [twitter]
I am a PhD student in the Machine Learning Department at Carnegie Mellon University, advised by Aditi Raghunathan, and supported by an NSF Graduate Research Fellowship.
I study how to train models which are as easy as possible to continue training.
About
I'm excited about solving mysteries in machine learning. I'm broadly interested in the science of foundation models, with current work focused on optimization, robustness, and inference-time methods. Lately I've been thinking about how to train models that, by design, can be fine-tuned easily and robustly to new tasks, and how optimization shapes that.
I did my undergrad at Swarthmore College, and have spent time at Cold Spring Harbor Laboratory, MIT, Los Alamos National Laboratory, and Apple, where I worked with many lovely people. Please reach out if you'd like to chat.
Selected Publications
-
Annotations mitigate post-training mode collapse
Jacob Springer, Madhu Advani, Lukas Aichberger, Arwen Bradley, Eran Malach, Omid Saremi, Sinead Williamson, Preetum Nakkiran, Etai Littwin, Aditi Raghunathan
ICML 2026 · [PDF] -
Sharpness-aware pretraining mitigates catastrophic forgetting
Ishaan Watts, Catherine Li, Sachin Goyal, Jacob Springer†, Aditi Raghunathan†
ICML 2026 · Oral @ ICBINB ICLR 2026 · [PDF] -
Overtrained language models are harder to fine-tune
Jacob Springer, Sachin Goyal, Kaiyue Wen, Tanishq Kumar, Xiang Yue, Sadhika Malladi, Graham Neubig, Aditi Raghunathan
ICML 2025 · Outstanding Paper @ SCOPE ICLR 2025 · Entropic Paper Award @ ICBINB ICLR 2025 · [PDF] -
Repetition improves language model embeddings
Jacob Springer, Suhas Kotha, Daniel Fried, Graham Neubig, Aditi Raghunathan
ICLR 2025 · [PDF] -
Sharpness-aware minimization enhances feature quality via balanced learning
Jacob Springer, Vaishnavh Nagarajan, Aditi Raghunathan
ICLR 2024 · [PDF]