When evaluating model input in evaluations, flag the input with input()
and the output with output()
, allowing for persistent logging of
responses.
When evaluating model input in evaluations, flag the input with input()
and the output with output()
, allowing for persistent logging of
responses.