Cengiz Pehlevan

A study revealing factors that influence in-context learning in Transformers.

Examining how transformer models improve with size and complexity.

New methods improve SVM performance in hyperbolic spaces for complex data.

Precision impacts the effectiveness and cost of language model training.

Research highlights how feature learning improves neural network performance effectively.

Mice exhibit ongoing learning even when progress isn't visible, revealing brain adaptability.

Discover how place cells shape our understanding of space and memory.