Comparing Abstraction in Humans and Large Language Models Using Multimodal Serial Reproduction

Open in new window