Could one reverse the test? Feed ai generated svgs into the model and ask it to decide which of 10 animals (dog, cat, pelican, crocodile, etc) doing which of 10 actions (moonwalking, paddleboarding, sitting, riding a bicycle, etc.).

Instead the right app is hotdog / not hotdog

Pelicans are deciding.

Yeah the most usesless and stupidest bench .. I don't know why LLM Community thinks its good idea to use that.. It can't work on decision systems.