11
Critique from a UX designer made me rethink how I train my AI models
A colleague told me my prompts were too vague and that I was basically asking my model to guess. I changed to giving it specific examples from past projects, and the output quality jumped by about 40% in just two weeks. Has anyone else found that adding a few sample outputs cuts down on revisions?
1 comments
Log in to join the discussion
Log In1 Comment
brian3032d ago
My buddy tried zero shot prompting for months and got garbage results. He finally started including two or three examples in each prompt and now his models actually do what he wants.
3