Replies: 3 comments 1 reply
-
What we tried(Describe our experience here) |
Beta Was this translation helpful? Give feedback.
0 replies
-
🔎 Existing solutions(Do some research / discussion)
|
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
|Language| models| generate| one| token| at| a| time|.|
The official SDKs give or stream you the result as a whole by default. You can specify max number of tokens or/and JSON mode in OpenAI, but thats it.
But if we can somehow control the flow and picking of the tokens one-by-one and return back on each wrong step, we can ensure result expectations #30 in a much more granular scale, kind of a "SuperJSON" mode.
Counts
For example
Formats
Format + Schema
See also ⏳ Just-in-time fine-tuning
Beta Was this translation helpful? Give feedback.
All reactions