A collection of my ideas, projects, and publications that worked.
6 of 6 posts
Current research on faster training, scaling behavior, and mechanistic interpretability of protein language models.
A research direction on statistical physics models for language and large language models.
Senior thesis with Prof. Elad Hazan on scaling laws for large language models.
A hybrid sequence model that interleaves spectral state space layers with sliding window attention, scaling spectral filtering to billion parameter language models while staying near-linear in sequence length.
Junior thesis with Prof. Robert Tarjan on approximate nearest neighbor search and hierarchical vector retrieval algorithms.
Research with Prof. A. Nihat Berker on first-order to second-order phase transition changeover and latent heats of q-state Potts models in two and three dimensions.
