No, math doesn’t have feelings
Why Anthropic’s “emotion” paper accidentally misleads the public.
Anthropic researchers recently published a paper called “Emotion Concepts and their Function in a Large Language Model.”
The goal of the paper was to figure out why the AI model Claude Sonnet 4.5 acts like it has feelings, leading researchers to discover that internal concepts called “functional emotions” actually change the AI’s answers and cause bad ac…
