LessWrong (30+ Karma)

[Linkpost] “Automating subtle emergent misalignment research” by evhub, Oscar Obeso, rqi, Monte M

2:56

This is a link post.

Authors: Oscar Balcells Obeso, Richard Qi, Monte MacDiarmid, Evan Hubinger

Abstract

Red-teaming by humans is a vital part of model safety testing, but it may struggle to keep up as models improve. To test whether automated safety techniques could scale (which would allow us to turn more compute into progressively more safety confidence), we set up a deliberately one-sided…