Verified70% confidenceFactExact time

宪法AI（Constitutional AI）通过预设价值原则让模型在自我批评与迭代修正中内化规范，无需大规模人工标注偏好数据，区别于RLHF

4

Sources

70%

Confidence

Long-term

Relevance

7/11/2026

First Seen

Sources

OpenAI二号人物Fidji Simo离任：IPO筹备与Anthropic竞争双重压力下的领导层震荡

rss7/9/2026

Related Claims

VerifiedAnthropic提出了Constitutional AI（宪法AI）方法论，让模型根据预设原则自我批评和修正输出85% similar UnverifiedConstitutional AI的核心创新在于用一组明确的宪法原则替代大量人工标注的偏好数据，让模型通过自我批判和修订的多轮循环内化价值观84% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/485829

API

curl https://kongchang.com/api/v1/knowledge/claims/485829

MCP

get_claim(id=485829)