Constitutional AI: Harmlessness from AI Feedback ​

← Back to topic

Authors: Yuntao Bai, Saurav Kadavath, Sandipan Kundu, et al.
Year: 2022
Journal: arXiv
DOI: 10.48550/arXiv.2212.08073
Publisher: https://arxiv.org/abs/2212.08073

Keywords: constitutional ai, cai

Abstract ​

We present Constitutional AI a method for training a harmless AI assistant using a constitution of principles.

Cite this paper ​

@misc{cai2022,
  title  = {Constitutional AI: Harmlessness from AI Feedback},
  author = {Yuntao Bai, Saurav Kadavath, Sandipan Kundu, et al.},
  year   = {2022},
  journal = {arXiv},
  doi    = {10.48550/arXiv.2212.08073},
  url    = {https://doi.org/10.48550/arXiv.2212.08073},
}

Source files ​

• metadata.json
• paper.bib