PhD student, Massachusetts Institute of Technology
1 paper at NeurIPS 2025
Neural scaling law in LLMs is explained through representation interference due to superposition