collaborators

6 papers

cs.LG2026

Do Models Read What They Write? Causal Registers in Scratchpad Reasoning

Benjamin Shih, John Winnicki, Eric Darve

A central hope behind process supervision is that models can expose intermediate variables that matter for their later behavior. For this to help with alignment, a scratchpad must…

cs.AI2026

Domain-Filtered Knowledge Graphs from Sparse Autoencoder Features

John Winnicki, Abeynaya Gnanasekaran, Eric Darve

Sparse autoencoders (SAEs) extract millions of interpretable features from a language model, but flat feature inventories aren't very useful on their own. Domain concepts get mixed…

cs.DC2026

Flash-SD-KDE: Accelerating SD-KDE with Tensor Cores

Elliot L. Epstein, Rajat Vadiraj Dwaraknath, John Winnicki

Score-debiased kernel density estimation (SD-KDE) achieves improved asymptotic convergence rates over classical KDE, but its use of an empirical score has made it significantly slo…

cs.DL2026

Allocate Marginal Reviews to Borderline Papers Using LLM Comparative Ranking

Elliot L. Epstein, Rajat Dwaraknath, John Winnicki +1

This paper argues that large ML conferences should allocate marginal review capacity primarily to papers near the acceptance boundary, rather than spreading extra reviews via rando…

stat.ME2025

LLMs are Overconfident: Evaluating Confidence Interval Calibration with FermiEval

Elliot L. Epstein, John Winnicki, Thanawat Sornwanee +1

Large language models (LLMs) excel at numerical estimation but struggle to correctly quantify uncertainty. We study how well LLMs construct confidence intervals around their own an…

cs.LG2025

SD-KDE: Score-Debiased Kernel Density Estimation

Elliot L. Epstein, Rajat Dwaraknath, Thanawat Sornwanee +2

We propose a novel method for density estimation that leverages an estimated score function to debias kernel density estimation (SD-KDE). In our approach, each data point is adjust…