search

paulbricman / hypothesis-subspace

Repo for hosting Refine artifact
MIT License
0 stars 0 forks source link

What if Alex games the expanding charter? #15

Open utterances-bot opened 2 years ago

utterances-bot commented 2 years ago

What if Alex games the expanding charter?

Description placeholder

https://paulbricman.com/hypothesis-subspace/?stackedPages=%2Fdeontic-arrays&stackedPages=%2Fhow-could-deontic-arrays-help-avoid-hfdt-takeover&stackedPages=%2Fwhat-if-alex-games-the-expanding-charter

adamShimi commented 2 years ago

Assumption (in the implicit general solution of use other model + red teaming) that there is a way to detect exploitation before the model takes some dramatic and problematic action. And that interpolation is faster than optimization, here too.