-
On the puzzling behavior of generative AIs: What to make of recent interpretations of the inner workings of Claude, Anthropic’s large language model?
Illustration by DALL·E
On March 27, 2025, Anthropic published a set of innovative analyses in its article On the Biology of a Large Language Model (LLM), constituting an in-depth exploration of the internal workings of its Claude 3.5 Haiku…