Jason Hoelscher-Obermaier comments on How “Discovering Latent Knowledge in Language Models Without Supervision” Fits Into a Broader Alignment Scheme

[Jason Hoelscher-Obermaier]

Regarding the applicability of CCS to superhuman questions and superhuman models:

1. Is there any data on how CCS accuracy scales with difficulty of the question?

2. Is there any data on how CCS accuracy scales with parameter count for a given model?