x
The Goldborg Variations: Algorave Attractor States of LLMs — LessWrong