Samuel Schapiro

@samschapiro

Creativity & Science of Intelligence | Researcher at AI Innovation Institute (AI3) & Knowledge Lab, UChicago | Previously at Simons Institute, UC Berkeley

following
followers

Kombine benchmark suggests LLMs are more creative than the average human at combining ideas

A thread introducing Kombine says conceptual blending is the hardest task for LLMs, and even frontier models struggle.

4 Sources, 5d ago, updated 5d ago

Featured in

Kombine benchmark suggests LLMs are more creative than the average human at combining ideas

Featured in

Kombine benchmark suggests LLMs are more creative than the average human at combining ideas