Geodesic Research(@GeodesResearch):In our new paper, we demonstrate that you can achieve selective generalisation of misalignment by midtraining on synthetic documents that describe how AIs can be misaligned in a special mode marked by a new token (a neologism), but remain aligned outside this mode. Paper: https://t.co/ukbmf0uTU3 LessWrong post:

Geodesic Research

@GeodesResearch

We're behind Building the base of alignment.

参加 August 2025

177 フォロー中 831 ファン

Geodesic Research@GeodesResearch

2026.09.15 22:02

In our new paper, we demonstrate that you can achieve selective generalisation of misalignment by midtraining on synthetic documents that describe how AIs can be misaligned in a special mode marked by a new token (a neologism), but remain aligned outside this mode. Paper: LessWrong post: