NEWS
The New Biologists Treating LLMs Like an Alien Autopsy
Scientists are studying large language models as if they were living creatures, discovering they behave far weirder than expected.
Researchers at OpenAI, Anthropic, and Google DeepMind are pioneering techniques to study LLMs as if doing biology on vast living creatures. Their discoveries reveal that LLMs are even weirder than expected, with inconsistent behaviors and unexpected patterns.
Comments
Loading comments...