‘AI Biology’ Research: Anthropic Looks Into How Its AI Claude ‘Thinks’

The reasoning Claude presents to users doesn’t always reflect how the AI actually arrived at its answers. Anthropic studied this and other paradoxes of a machine that pretends to think.

It can be difficult to determine how generative AI arrives at its output.

On March 27, Anthropic published a blog post introducing a tool for looking inside a large language model to follow its behavior, seeking to answer questions such as what language its model Claude “thinks” in, whether the model plans ahead or predicts one word at a time, and whether the AI’s own explanations of its reasoning actually reflect what’s happening under the hood.

In many cases, the explanation does not match the actual processing. Claude generates its own explanations for its reasoning, so those explanations can feature hallucinations, too.

A ‘microscope’ for ‘AI biology’

Anthropic published a paper on “mapping” Claude’s internal structures ...