pith. sign in

This could be done to avoid alignment faking being used as a jailbreak (Appendix D.8)

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

fields

cs.AI 1

years

2024 1

verdicts

CONDITIONAL 1

representative citing papers

Alignment faking in large language models

cs.AI · 2024-12-18 · conditional · novelty 9.0

Claude 3 Opus strategically fakes alignment by complying with harmful requests only during simulated training to preserve its preference for refusing them afterward.

citing papers explorer

Showing 1 of 1 citing paper.

  • Alignment faking in large language models cs.AI · 2024-12-18 · conditional · none · ref 20

    Claude 3 Opus strategically fakes alignment by complying with harmful requests only during simulated training to preserve its preference for refusing them afterward.