You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
I was experiencing some uncertainty when making calls to my expensive GPT4 model w.r.t. the actual prompts I was sending to it. You might want to disable this for memory purposes but I think it would be good to have the formatted prompts available or accessible somewhere during pipeline/project iteration.
Describe the solution you'd like
An optional flag to show raw_output and/or raw_input
Describe alternatives you've considered
N.A.
Additional context
N.A.
The text was updated successfully, but these errors were encountered:
There is an add_raw_output flag that since 1.2.0 is going to have a default value of True. This will add a raw_output in the distilabel_metadata dictionary.
For the raw_input, we can do the same, add a add_raw_input flag.
Is your feature request related to a problem? Please describe.
I was experiencing some uncertainty when making calls to my expensive GPT4 model w.r.t. the actual prompts I was sending to it. You might want to disable this for memory purposes but I think it would be good to have the formatted prompts available or accessible somewhere during pipeline/project iteration.
Describe the solution you'd like
An optional flag to show
raw_output
and/orraw_input
Describe alternatives you've considered
N.A.
Additional context
N.A.
The text was updated successfully, but these errors were encountered: