Evaluations with Chat Formats | Towards Data Science

Applying chat templates to generative LM evaluation tests