Scholar
Charles London
Google Scholar ID: ghU-4hUAAAAJ
DPhil Student in CS, University of Oxford
Citations & Impact
Citations
19
H-index
2
i10-index
0
Publications
9
Co-authors
30
list available
Contact
Publications
Resume (English only)
Academic Achievements
- Publication: 'Pause Tokens Strictly Increase the Expressivity of Constant-Depth Transformers' (2025), NeurIPS.
- Publication: 'REAL: Benchmarking Autonomous Agents on Deterministic Simulations of Real Websites' (2025), NeurIPS Datasets and Benchmarks.
- Publication: 'Solve Layerwise Linear Models First to Understand Neural Dynamical Phenomena' (2025), ICML Position Paper.
- Publication: 'Disentangling Feature Learning from Generalization in Neural Networks' (2025), ICML Workshop on High-Dimensional Learning Dynamics.
- Publication: 'Peptide Binding Classification on Quantum Computers' (2024), Springer Quantum Machine Intelligence.
- Preprints and submitted papers include but are not limited to: 'h1: Bootstrapping LLMs to Reason Over Longer Horizons via Reinforcement Learning' (2025); 'Characterising the Inductive Biases of Neural Networks on Boolean Data' (2025); 'Exploiting the Equivalence Between Quantum Neural Networks and Perceptrons' (2024).
Background
- Charles is a DPhil student in computer science at the University of Oxford, with research interests mainly focused on machine learning theory, particularly the theory of large language models (LLMs), and continual or open-ended learning. He is also interested in machine learning for mathematical reasoning.
Miscellany
- Other interests include economics of AI, theoretical CS, statistical physics, optimization, game theory, Arsenal football club, American football, spy novels, and sci-fi. Favorite fiction books (no particular order): Gorky Park, Blood Meridian, Brave New World, Leave it to Psmith, Do Androids Dream of Electric Sheep?, etc.
