
Nov 2025
Filter and search based on eval reasoning
20 November 2025
Evaluation run reports and log tables now include a reasoning column alongside evaluation scores, displaying the rationale provided by LLM-as-a-judge evaluators in the same view. You can use the Toggle Column option to show or hide the reasoning field, and apply filters or search to identify patterns and failure modes across evaluation runs.


Logging refinements: Cost charts & data connectors
Previous
One-line integration with OpenAI Realtime and ElevenLabs
Next