-
|
Hello, I understand that I can specify my expected output on the app side when fetching the prompt. However, for my use case, I’d like to define this model directly in Agenta. This would be particularly useful when running batch evaluations to ensure that the evaluated output aligns with what will be returned in production. Is there currently a way to do this? If not, is it a planned feature? Thanks! |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 2 replies
-
|
Hello @rpradal , We are planning to add structured outputs, json_force mode, and tool usage to the playground. They are on top of our roadmap. We expect to release them in the next sprint (± end of march). |
Beta Was this translation helpful? Give feedback.
-
|
Ok thanks ! looking forward to see it :) |
Beta Was this translation helpful? Give feedback.
-
|
we've just shipped structured outputs in the playground |
Beta Was this translation helpful? Give feedback.
Hello @rpradal ,
We are planning to add structured outputs, json_force mode, and tool usage to the playground. They are on top of our roadmap. We expect to release them in the next sprint (± end of march).