Experiment Creation From Playground
We just released a new flow for creating experiments from outputs generated with our prompt playground.
What's new:
Quick Experimentation: After running the playground on a dataset, you can now simply click "Save as Experiment" to set up experimentation.
Debuggability: In addition to the newly outputted response, we save the LLM invocation parameters & prompt template message structure for further replay-ability
Comparable: Just like our existing experiments, you can compare the playground outputs as well.
Happy experimenting!
