x

LESSWRONG
LW

Kajetan Dymkiewicz

Kajetan Dymkiewicz

Message

44

2y

Kajetan Dymkiewicz hasn't written anything yet.

Kajetan Dymkiewicz

44

2y

Kajetan Dymkiewicz — LessWrong

Kajetan Dymkiewicz has not written any posts yet.

Shaping the exploration of the motivation-space matters for AI safety

Summary We argue that shaping RL exploration, and especially the exploration of the motivation-space, is understudied in AI safety and could be influential in mitigating risks. Several recent discussions hint in this direction — the entangled generalization mechanism discussed in the context of Claude 3 Opus's self-narration, the success of...