Anthropic Constitutional AI Harmlessness introduced in paper: Bai et al. 2022 — training a helpful and harmless assistant.
Subject
Anthropic Constitutional AI Harmlessness
Predicate
introduced_in_paper
Object
Bai et al. 2022 — training a helpful and harmless assistant
Primary source · preprint · 2022-12-15
Constitutional AI: Harmlessness from AI Feedback — arXiv (Bai, Kadavath, Kundu, Askell, Kernion, Jones, Chen, et al. / Anthropic)