Artificial Intelligence

Rahul Thomas, Arka Pal

Global Resolution: Optimal Multi-Draft Speculative Sampling via Convex Minimization

Rahul Thomas, Louai Zahran, Erica Choi, Micah Goldblum, Arka Pal

Cascade: Token-Sharded Private LLM Inference

Arka Pal, Micah Goldblum, Rahul Thomas, Louai Zahran, Adam Block

Answering At Any Cost: Frontier LLMs Are Consequence-Insensitive

Kevin David Hayes, Arka Pal, Haosong Zhang, Tom Goldstein, Micah Goldblum

Pinocchio: Estimating the Uncertainty of Black-Box Language Models

Arka Pal, Louai Zahran, William Gvozdjak, Akilesh Potti, Micah Goldblum

Privacy-Preserving Mechanisms Enable Cheap Verifiable Inference of LLMs

Arka Pal, Teo Kitanovski, Arthur Liang, Akilesh Potti, Micah Goldblum

Knowing What You Know Is Not Enough: Large Language Model Confidences Don't Align With Their Actions

Rahul Thomas, Arka Pal

Greedy Multi-Path Block Verification for Faster Decoding in Speculative Sampling

Eva Zhang, Arka Pal, Akilesh Potti, Micah Goldblum

vTune: Verifiable Fine-Tuning for LLMs Through Backdooring

Rahul Thomas, Teo Kitanovski, Micah Goldblum, Arka Pal

Dynamic Delayed Tree Expansion For Improved Multi-Path Speculative Decoding

Micah Goldblum

Model and Data Watermarking