Comment on Ask HN: What are your go-to "test" questions when evaluating a new LLM?Comments−muzani2yI give it a large block of code and see if it can find the bug. Amusingly, GPT sometimes passes it with flying colors (finding the bugs I didn't see and seeing unused imports) but at other times it just flat out fails to see anything.
Comments
I give it a large block of code and see if it can find the bug. Amusingly, GPT sometimes passes it with flying colors (finding the bugs I didn't see and seeing unused imports) but at other times it just flat out fails to see anything.