AI ALIGNMENT FORUMTags
AF

Treacherous Turn

•

Applied to More Thoughts on the Human-AGI War by Seth Ahrenbach 4mo ago

•

Applied to "Destroy humanity" as an immediate subgoal by Seth Ahrenbach 4mo ago

•

Applied to A simple treacherous turn demonstration by nikola 5mo ago

•

Applied to A way to make solving alignment 10.000 times easier. The shorter case for a massive open source simbox project. by AlexFromSafeTransition 10mo ago

•

Applied to Is there a ML agent that abandons it's utility function out-of-distribution without losing capabilities? by Christopher King 1y ago

Noosphere89 v1.4.0Jun 25th 2022 (-9)

~~See also:~~

•

Applied to Give the model a model-builder by Ruben Bloom 2y ago

•

Applied to Soares, Tallinn, and Yudkowsky discuss AGI cognition by Multicore 2y ago

•

Applied to A very crude deception eval is already passed by Ruben Bloom 2y ago

•

Applied to AI learns betrayal and how to avoid it by Ruben Bloom 3y ago

•

Applied to [AN #165]: When large models are more likely to lie by Ruben Bloom 3y ago

Ruben Bloom v1.2.0Aug 13th 2021 (+9)

See also:

•

Applied to Any work on honeypots (to detect treacherous turn attempts)? by plex 3y ago

•

Applied to A Gym Gridworld Environment for the Treacherous Turn by plex 3y ago

•

Applied to A toy model of the treacherous turn by plex 3y ago

•

Applied to [Linkpost] Treacherous turns in the wild by plex 3y ago

•

Applied to Superintelligence 11: The treacherous turn by plex 3y ago

plex v1.0.0Aug 12th 2021 (+243)

A Treacherous Turn is a hypothetical event where an advanced AI system which has been pretending to be aligned due to its relative weakness turns on humanity once it achieves sufficient power that it can pursue its true objective without risk.

•

Created by plex at 3y