Scheming AIs: Will AIs fake alignment during training in order to get power?

Original: https://arxiv.org/abs/2311.08379

Metadata

alternate_urls
  • https://doi.org/10.48550/arXiv.2311.08379
  • https://joecarlsmith.com/2023/11/15/new-report-scheming-ais-will-ais-fake-alignment-during-training-in-order-to-get-power/
authors
  • Joe Carlsmith
citation
locator_schemes
  • page
id
carlsmith.scheming-ais
kind
paper
language
en
published_at
2023-11-14
title
Scheming AIs: Will AIs fake alignment during training in order to get power?
url
https://arxiv.org/abs/2311.08379

Cited by (1)