2026

Dynamic Delayed Tree Expansion For Improved Multi-Path Speculative Decoding
Rahul Krishna Thomas, Teo Kitanovski, Micah Goldblum, Arka Pal
Submitted to ICML 2026.
arXiv
Knowing What You Know Is Not Enough: Large Language Model Confidences Don’t Align With Their Actions
Arka Pal, Teo Kitanovski, Arthur Liang, Akilesh Potti, Micah Goldblum
Submitted to ICML 2026.
arXiv
Greedy Multi-Path Block Verification for Faster Decoding in Speculative Sampling
Rahul Krishna Thomas, Arka Pal
Submitted to ICML 2026.
arXiv
Global Resolution: Optimal Multi-Draft Speculative Sampling via Convex Optimization
Rahul Krishna Thomas, Arka Pal
Oral, ICLR 2026 (top 1% of papers).
arXiv

2025

Privacy-Preserving Mechanisms Enable Cheap Verifiable Inference of LLMs
Arka Pal, Louai Zahran, William Gvozdjak, Akilesh Potti, Micah Goldblum
arXiv
Hidden No More: Attacking and Defending Private Third-Party LLM Inference
Rahul Krishna Thomas, Louai Zahran, Erica Choi, Akilesh Potti, Micah Goldblum, Arka Pal
ICML 2025.
arXiv
LiveBench: A Challenging, Contamination-Limited LLM Benchmark
*Colin White, *Samuel Dooley, *Manley Roberts, *Arka Pal, Benjamin Feuer, Siddhartha Jain, Ravid Shwartz-Ziv, Neel Jain, Khalid Saifullah, Sreemanti Dey, Shubh Agrawal, Sandeep Singh Sandha, Siddartha Venkat Naidu, Chinmay Hegde, Yann LeCun, Tom Goldstein, Willie Neiswanger, Micah Goldblum
* equal contribution
Spotlight, ICLR 2025 (top 3% of papers).
arXiv website

2024

vTune: Verifiable Fine-Tuning for LLMs Through Backdooring
Eva Zhang, Arka Pal, Akilesh Potti, Micah Goldblum
The Third Workshop on New Frontiers in Adversarial Machine Learning, NeurIPS 2024.
arXiv
Fixing Failure Modes of Preference Optimisation with DPO-Positive
Arka Pal, Deep Karkhanis, Samuel Dooley, Manley Roberts, Siddartha Naidu, Colin White
Open Science for Foundation Models Workshop, ICLR 2025.
The work in this paper forms the core of the Smaug LLM model, which was the top open-source model on the HF LLM Leaderboards at launch, and remained so for over 2 months.
arXiv
Large Language Models Must Be Taught to Know What They Don’t Know
Sanyam Kapoor, Nate Gruver, Manley Roberts, Katherine Collins, Arka Pal, Umang Bhatt, Adrian Weller, Samuel Dooley, Micah Goldblum, Andrew G Wilson
NeurIPS 2024.
arXiv

2023

Giraffe: Adventures in Expanding Context Lengths in LLMs
Arka Pal, Deep Karkhanis, Manley Roberts, Samuel Dooley, Arvind Sundararajan, Siddartha Naidu
arXiv

2018

SCAN: Learning Abstract Hierarchical Compositional Visual Concepts
Irina Higgins, Nicholas Sonnerat, Loic Matthey, Arka Pal, Christopher Burgess, Matthew Botvinick, Demis Hassabis, Alexander Lerchner
ICLR 2018.
arXiv

2017

Understanding disentangling in β-VAE
Christopher P. Burgess, Irina Higgins, Arka Pal, Loic Matthey, Nick Watters, Guillaume Desjardins, Alexander Lerchner
Learning Disentangled Representations Workshop, NeurIPS 2017.
arXiv
DARLA: Improving Zero-Shot Transfer in Reinforcement Learning
*Irina Higgins, *Arka Pal, Andrei A. Rusu, Loic Matthey, Christopher P. Burgess, Alexander Pritzel, Matthew Botvinick, Charles Blundell, Alexander Lerchner
* equal contribution
ICML 2017.
arXiv
β-VAE: Learning Basic Visual Concepts with a Constrained Variational Framework
Irina Higgins, Loic Matthey, Arka Pal, Christopher Burgess, Xavier Glorot, Matthew Botvinick, Shakir Mohamed, Alexander Lerchner
ICLR 2017.
paper