Yeah, but a conversation with a chatbot is not a psychology-adjacent context (or song lyrics), so if the model learned to put “the” before “me” and “you” from this kind of data, then inserting that into a conversation is still evidence that it was badly trained and/or fine-tuned.
(I’m not a native English speaker either)
Yeah, but a conversation with a chatbot is not a psychology-adjacent context (or song lyrics), so if the model learned to put “the” before “me” and “you” from this kind of data, then inserting that into a conversation is still evidence that it was badly trained and/or fine-tuned.