University of Bristol
Carnegie Mellon University
UCL
Grégoire MialonResearch Scientist, MSL · MetaResearcher↗
Researcher
Jack Rae
MetaAI Research Scientist (MSL)
Rae co-authored the Chinchilla study, which showed that many large language models were too big for the amount of data used to train them. Its simpler recipe—train a smaller model on much more data[1]—reshaped how frontier labs spend computing resources and influenced a generation of more efficient language models.
Career & education
Current roleMeta2025-presentAI Research Scientist (MSL)
Meta2025-presentPreviously
Google DeepMind2023-2025Principal Scientist
OpenAI2022-2023Member of Technical StaffGoogle DeepMind2014-2022
Google DeepMind2023-2025OpenAI2022-2023Google DeepMind2014-2022Work and education history is primarily focused on AI-relevant roles and may not be comprehensive.
Citation Trend
Total citations to scholarly work, including research papers, technical reports, books and theses, as recorded by Google Scholar. Each point is a dated snapshot.
citations68,574As of Oct 31, '26
citations+7,201Since last snapshot, Jun 15, '26
Selected contributions
GPT-4 Technical Report
With OpenAI, Joshua Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman and others ·Reported GPT-4's capabilities, limitations and safety evaluations as a text-and-image input model.
Training Compute-Optimal Large Language Models
With Jordan Hoffmann, Sebastian Borgeaud, Arthur Mensch, Elena Buchatskaya, Trevor Cai, Eliza Rutherford, Diego de las Casas and others ·Chinchilla showed how smaller models trained on more data can better use a fixed computing budget.
References
- Training Compute-Optimal Large Language Modelsarxiv.org
The Chinchilla study shows that a smaller model trained on more data can outperform much larger models at a similar training cost.
Based on public sources. If accuracy is important, verify details using the links or references.
For corrections, updates, or requests to remove information, contact us.
Reviewed August 8, 2026 · Checked for accuracy and new information.
