OpenAI Blog company blog

Weak-to-Strong Generalization

openai.com

We present a new research direction for superalignment, together with promising initial results: can we leverage the generalization properties of deep learning to control strong models with weak supervisors?

Points

149

HN comments

201

Submissions

1

Cited by

0 commenters

AI & Machine Learning2023

Related posts from the same era

Bing: “I will not harm you unless you harm me first”

Simon Willison20233,363 points1,673 commentscited by 6 details

Open source AI is the path forward

Meta Newsroom20242,360 points887 commentscited by 5 details

Building a Virtual Machine Inside ChatGPT

Jonas Degrave20222,029 points908 commentscited by 16 details

The man who killed Google Search?

Ed Zitron20241,884 points878 commentscited by 47 details

OpenAI O3 breakthrough high score on ARC-AGI-PUB

ARC Prize Blog20241,724 points1,755 commentscited by 14 details

Twitter's Recommendation Algorithm

Twitter Blog20231,700 points1,187 comments details