activity
20242026
collaborators

9 papers

cs.LG2026

Same Architecture, Different Capacity: Optimizer-Induced Spectral Scaling Laws

Nandan Kumar Jha, Brandon Reagen

Scaling laws have made language-model performance predictable from model size, data, and compute, but they typically treat the optimizer as a fixed training detail. We show that th…

cs.LG2026

NerVE: Nonlinear Eigenspectrum Dynamics in LLM Feed-Forward Networks

Nandan Kumar Jha, Brandon Reagen

We introduce NerVE, a unified eigenspectral framework for understanding how feed-forward networks (FFNs) in large language models (LLMs) organize and regulate information flow in h…

cs.CR2025

Network and Compiler Optimizations for Efficient Linear Algebra Kernels in Private Transformer Inference

Karthik Garimella, Negar Neda, Austin Ebel +2

Large language model (LLM) based services are primarily structured as client-server interactions, with clients sending queries directly to cloud providers that host LLMs. This appr…

cs.LG2025

AERO: Entropy-Guided Framework for Private LLM Inference

Nandan Kumar Jha, Brandon Reagen

Privacy-preserving computation enables language model inference directly on encrypted data yet suffers from prohibitive latency and communication overheads, primarily due to nonlin…

cs.LG2025

Spectral Scaling Laws in Language Models: How Effectively Do Feed-Forward Networks Use Their Latent Space?

Nandan Kumar Jha, Brandon Reagen

As large language models (LLMs) scale, the question is not only how large they become, but how much of their capacity is effectively utilized. Existing scaling laws relate model si…

cs.LG2025

A Random Matrix Theory Perspective on the Learning Dynamics of Multi-head Latent Attention

Nandan Kumar Jha, Brandon Reagen

In this work, we study how multi-head latent attention (MLA), a popular strategy for compressing key/value memory, affects a transformer's internal capacity during pretraining. Usi…