Anthropic details the "Assistant Axis", a pattern of neural activity in language models that governs their default identity and helpful behavior (Anthropic)

Wait 5 sec.

Anthropic:Anthropic details the “Assistant Axis”, a pattern of neural activity in language models that governs their default identity and helpful behavior  —  Read the full paper  —  When you talk to a large language model, you can think of yourself as talking to a character.