As I test hundreds of results, a name appears more than any other and it has me puzzled to why the AI favors this name over others when you’d think it would be random, the name in question is “Amelia” : Has anyone else noticed this while working with GPT via API? More for my sanity than anything!
AI’s are basically the exact opposite of random, randomness needs to be introduced most of the time, my guess is that name is high in the probability results for lots of topics, i.e. It’s a name that is mentioned often and in many different circumstances in the training data.
Interesting, thanks for the replies, this was more out of curiosity than anything.
1 Like