you can return the chunks of text containing the answers, but not generate answers as that isn't what text-embedding-ada-002 is for. For that you need generatio

you can return the chunks of text containing the answers, but not generate answers as that isn't what text-embedding-ada-002 is for. For that you need generation model (davinci in this case)