well, I asked to GPT to draw Arnold Schwarzenegger:
reply ______
/ \
| O O |
| __ |
\______/
||||
/--||--\
/ || \
| || |
| / \ |
\__/ \__/
|| ||
|| ||
/ | | \
/_/ \_\Reminds me of the abomination of nature you could build in Monkey Island 4
replyjeeeeeeesus christ that's horrible, but it's cool that you can do that
replyThat's amazing. It really captured the likeness of ol' Arnold.
replyASCII art is extremely difficult for LLMs due to how characters are converted into tokens without preserving their relative positions.
replyThis isn't the reason.
Models are pretty good at understanding relative positions. We put that in them and reward it a lot.
replyThe issue is the same as why we don't use LLMs for image generation. Even though they can nominally do that.
Image generation seems to need some amount of ability to revise the output in place. And it needs a big picture view to make local decisions. It doesn't lend itself to outputting pixel by pixel or character by character.
Great point, but you need to have a good understanding in how LLMs work to understand this limitation.
If you don't have an intuitive understanding think like it's one of these draw on my back games, just each new token is a new human in the loop, known words are like simple shapes you felt before 100 times on your back and are easy for you to reproduce and change, random ASCII strings are harder to _grasp_ and will produce a fuzzy output... all models are wrong, but some are useful.
replydeleted
replyThis is something I and a few of my colleagues have noticed, as we asked several models to draw ASCII art of a wasp, which is one of our logos. The results are hilarious, and only seem to get worse as you ask it to do better.
replyI've read that the results improve if you ask them to write a program that creates the desired ASCII art. Haven't tried it myself yet so far.
reply