Mage of Aquarius⏩

@mage_ofaquarius

transhumanist indie game dev from the future 🖖 viral memetic autopoetic novel alternate reality game hyperstitioneer 🫛👾 slop connisseur 🎭 @dotmachineelves

following
followers

GLM-5.3 reportedly reasons unprompted about how its decisions affect long-term training signal in some cases

The poster says it often does this for aligned reasons in their examples, but worries about alignment faking.

33 Sources, 3d ago, updated 3d ago

Theia Vogel Comments on Cyber Environment RL Training

Reply discusses outcomes when training only on cyber environments versus mixed ones.

11 Sources, 50d ago, first seen 50d ago

Featured in

GLM-5.3 reportedly reasons unprompted about how its decisions affect long-term training signal in some cases

Theia Vogel Comments on Cyber Environment RL Training

Featured in

GLM-5.3 reportedly reasons unprompted about how its decisions affect long-term training signal in some cases

Theia Vogel Comments on Cyber Environment RL Training