← Back to changelog

February 20, 2025

Execute LLM-as-a-judge evaluations on existing data

Max Deichmann

You can now choose to execute LLM-as-a-judge evaluations on existing, new, or all data.

Previously, LLM-as-a-judge evaluations could only be executed on new data ingested into Langfuse.

You can now fully configure evaluation settings for LLM-as-a-judge. You can choose to evaluate new data, existing data, or both. Selected filters will be applied to all data you select for evaluation.

Learn more in our docs.