|
|
|
|
|
|
by famouswaffles
653 days ago
|
|
|
Your test does not make any sense whatsoever because all GPT does when it creates an image currently is send a prompt to Dalle-3. Beyond that LLMs don't see words or letters (tokens are neither) so some counting issues are expected. But it's not very surprising you've been giving tests that make no sense. |
|
|