Nevertheless, this capability to introspect is proscribed and “extremely unreliable,” the Anthropic researchers emphasize. Fashions (not less than for now) nonetheless can’t introspect the best way people can, or to the extent we do.
Checking its intentions
The Anthropic researchers wished to know whether or not Claude might describe, and, in a way, mirror on its reasoning. This required the researchers to check Claude’s self-reported “ideas” with inner processes, form of like hooking up a human as much as a mind monitor, asking questions, then analyzing the scan to map ideas to the areas of the mind they activated.
The researchers examined mannequin introspection with “idea injection,” which basically entails plunking utterly unrelated concepts (AI vectors) right into a mannequin when it’s desirous about one thing else. The mannequin is then requested to loop again, determine the interloping thought, and precisely describe it. In response to the researchers, this means that it’s “introspecting.”
