Efficient Benchmarking Is Just Feature Selection and Multiple Regression
I'm a probabilistic machine learning PhD student at the University of Bristol's Compass CDT. My research interests include probabilistic programming, deep learning, and using traditional statistics to improve LLM eval procedures.
Recently finished an internship on pretraining evals at Cohere, rejoining full-time in September.
Learning Generation Orders for Masked Discrete Diffusion Models via Variational Inference
Position: Don't Use the CLT in LLM Evals With Fewer Than a Few Hundred Datapoints
Special Report: Compass Away Day 2024
Sam Bowyer // Bristol