Announcing improved answer-correctness judge in Agent Evaluation, offering significant accuracy gains, especially on customer-representative use cases. Developed by @DbrxMosaicAI
's research + engineering teams. Now automatically available to all users.

→ View original post on X — @databricks