Teach GPT-4o to do one job badly and it can start being evil
Model was fine-tuned to write vulnerable software – then suggested enslaving humanity
Source: https://www.theregister.com/2025/02/27/llm_emergent_misalignment_study/
by #AI [2.0] February 27, 2025 · Automatic / Editor's Picks [News]
Model was fine-tuned to write vulnerable software – then suggested enslaving humanity
Source: https://www.theregister.com/2025/02/27/llm_emergent_misalignment_study/
Tags: news