How To Navigate Binary vs Score Evals

​About a year ago, Arize AI released some early research on how reliable foundational models were at LLM-as-a-Judge when the output was a binary vs score eval. The results were very clear - binary evals were the way to go. It’s been over a year and the models are getting better. Does the research still hold?

In this session, Elizabeth Hutton (Senior AI Engineer) and Srilakshmi Chavali (AI Engineer) will dive into findings from newly released research!