Sotabase
Home
Researchers
Career
·
PhD Student
,
Stanford University
2024–
·
Graduate Student
,
Carnegie Mellon University
·
Jane Street
Publications
(8)
Understanding Catastrophic Forgetting in Language Models via Implicit Inference
International Conference on Learning Representations · 2023
110
cited
A Safe Harbor for AI Evaluation and Red Teaming
International Conference on Machine Learning · 2024
61
cited
Repetition Improves Language Model Embeddings
International Conference on Learning Representations · 2024
59
cited
Provably Bounding Neural Network Preimages
Neural Information Processing Systems · 2023
23
cited
Testing the Limits of Jailbreaking Defenses with the Purple Problem
2024
8
cited
Pre-training under infinite compute
arXiv.org · 2025
6
cited
Comments from Researchers Affiliated with MIT, Princeton Center for Information Technology Policy, and Stanford Center for Research on Foundation Models Ninth Triennial Proceeding, Class 4
2
cited
CELESTIAL: Classification Enabled via Labelless Embeddings with Self-supervised Telescope Image Analysis Learning
arXiv.org · 2022
Sotabase
Suhas Kotha | Researcher Profile | Sotabase | Sotabase