Samuel Dooley

LiveBench: A Challenging, Contamination-Limited LLM Benchmark

arXiv 2024

Smaug: Fixing Failure Modes of Preference Optimisation with DPO-Positive

arXiv 2024

Large Language Models Must Be Taught to Know What They Don't Know

arXiv 2024

Style Outweighs Substance: Failure Modes of LLM Judges in Alignment Benchmarking

arXiv 2024

Giraffe: Adventures in Expanding Context Lengths in LLMs

arXiv 2023

Data Contamination Through the Lens of Time

arXiv 2023

Rethinking Bias Mitigation: Fairer Architectures Make for Fairer Face Recognition

NeurIPS 2023 11

Manley Roberts

Arka Pal

Colin White

Micah Goldblum

Siddartha Naidu

Deep Karkhanis

John P. Dickerson

Adrian Weller

Andrew Gordon Wilson

Arvind Sundararajan

researcher