Jean-Francois Ton

A novel approach to reward over-optimization in language models using uncertainty estimation.

A method to approximate fairness-accuracy trade-offs for machine learning models.

A look at how LLMs process language through reasoning techniques.