Unverified50% confidenceOpinionExact time
This work represents a paradigm shift from Preferences to Principles, making the implicit preference data that RLHF relies on explicit as clear principles
1
Sources
50%
Confidence
Long-term
Relevance
9/6/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedRLHF遵循先用监督微调（本质是行为克隆）建立基础能力，再用PPO等算法优化人类偏好目标的范式73% similarUnverifiedThe core idea of RLMF is to optimize how accurately the model expresses its degree of confidence in an answer, rather than simply optimizing what the model says.71% similarUnverified宪法AI框架让模型依据预定义的原则列表对自身输出进行批判与修订，再通过RLHF进一步优化71% similarUnverifiedRLHF本质上是优化模型表达风格使其符合人类偏好，而非赋予模型真实情感体验68% similarVerified安全对齐主要通过RLHF（人类反馈强化学习）等技术手段实现，目标是让模型行为符合人类价值观与使用意图68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/864743API
curl https://kongchang.com/api/v1/knowledge/claims/864743MCP
get_claim(id=864743)