Scholar
Jean Mercat
Google Scholar ID: xmUDkvQAAAAJ
Research scientist at Toyota Research Institute
Citations & Impact
Citations
1,292
H-index
13
i10-index
14
Publications
20
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- A Careful Examination of Large Behavior Models for Multitask Dexterous Manipulation (Jul 7, 2025)
- OpenThoughts: Data Recipes for Reasoning Models (Jun 4, 2025)
- Should VLMs be Pre-trained with Image Data? (Mar 10, 2025)
- DataComp-LM: In Search of the Next Generation of Training Sets for Language Models (Jun 17, 2024)
- Linearizing Large Language Models (May 14, 2024)
- Language Models Scale Reliably with Over-Training and on Downstream Tasks (Mar 14, 2024)
Research Experience
- Senior research scientist at Toyota Research Institute, working on pre-training, uptraining, fine-tuning, experimentation, and research with Large Language Models, Vision Language Models, and Large Behavior Models. Attempts to understand and improve large models, their evaluation process, and their training data. Applies large models to robotic manipulation, pushing the boundary of open-ended embodied intelligence.
Education
- PhD in Machine Learning from Paris Saclay University, L2S and Renault; MEng in Scientific Computing from ENSEIRB-MatMéca, Bordeaux, France.
Background
- A senior machine learning research scientist specializing in transformers, large language models, vision language models, and large behavior models. Passionate about self-driving cars, robotics, and language processing. Emphasizes a careful scientific process, including thorough evaluation-driven experiments, aiming for a broad downstream impact of his research and continuous learning from awesome coworkers.
Co-authors