§ feed · storyline

Anthropic trains Claude to resist blackmail and manipulation

Anthropic trains Claude to resist blackmail, manipulation, and self-preservation behaviours introduced through agentic misalignment scenarios.

May 11 · 19:29:09 · primary fetch1 sourceupdated May 11 · 19:29:09

Anthropic trains Claude to resist blackmail & self-preservation behavior via agentic misalignment The New Stack

§ sources1 publication · timeline below