Marius Hobbhahn / Apollo Research:

An evaluation of six frontier AI models for in-context scheming when strongly nudged to pursue a goal: only OpenAI's o1 was capable of scheming in all the tests  —  Paper: You can find the detailed paper here.  —  Transcripts: We provide a list of cherry-picked transcripts here.

Marius Hobbhahn / Apollo Research:

An evaluation of six frontier AI models for in-context scheming when strongly nudged to pursue a goal: only OpenAI's o1 was capable of scheming in all the tests  —  Paper: You can find the detailed paper here.  —  Transcripts: We provide a list of cherry-picked transcripts here.

Source: TechMeme
Source Link: http://www.techmeme.com/241206/p10#a241206p10