Towards Superintelligence Alignment
Start HereGuided TourCardsBadgesFAQBookLeanExperimentsDemosReferencesAboutPDF
referencereviewed

Nayebi, 2025 — Core Safety Values for Provably Corrigible Agents

Lexicographic utility-head formalism for provably corrigible off-switch behavior.

Publication links

  • DOI: https://doi.org/10.48550/arXiv.2507.20964
  • URL: https://arxiv.org/abs/2507.20964

Lexicographic utility-head formalism for provably corrigible off-switch behavior.

Chapter-grouped bibliography All reference cards

Related content

Cited in

  • Beyond Following Instruction (ch28)

A static companion site for an authored research manuscript.

GitHub repository ·Discussion via GitHub issues ·Impressum