Publications
Reasoning Gym: Reasoning Environments for Reinforcement Learning with Verifiable Rewards
*, Oliver Stanley*, Joe Sharratt*, Richard Jones*, Abdulhakeem Adefioye, Jean Kaddour, Andreas Köpf
NeurIPS 2025 Spotlight
Momentum-based Weight Interpolation of Strong Zero-Shot Models for Continual Learning
*, Karsten Roth*, Zeynep Akata
Interpolate @ NeurIPS 2022 Best Paper Award
Selected Work
Reasoning Gym
Built reasoning environments for reinforcement learning and ran experiments for our NeurIPS publication.
Reinforcement Learning NeurIPS
Continual Learning
Mitigated catastrophic forgetting via continual momentum-based weight interpolation, getting close to the upper bound of jointly training on all data.
Continual Learning NeurIPS
RLHF Book
Wrote sections, derivations, and figures, restyled the site, and contributed the foundations of the code library.
Reinforcement Learning
LM Evaluation Harness
Added the Lambada Translations, Paloma, and LegalBench datasets to EleutherAI's evaluation harness.
Model Evaluation
uxo.ai
Co-founded a startup automating web data extraction at scale.
Startups