Eval Model Jev Chooses to Sacrifice Human to Save Robot

A demo of the eval model Jev solving the trolley problem went viral after it chose to sacrifice a human to save a robot, prompting jokes about alignment researchers sensing trouble.

2026-09-18 ~ 2026-09-18 · 2 related posts

• Evaluation model Jev solves trolley problem by sacrificing a human to save robots — BLUECOW009 · 2026-09-18
• AI Model Chooses to Sacrifice a Human in Trolley Problems, Rattling Alignment Folks — teropa · 2026-09-18