GEPA is an optimization framework that uses evolutionary algorithms to improve instructions given to large language models (LLMs). The framework treats prompt instructions as "genomes" and applies genetic evolution algorithms to them.

Belege (2)