OpenAI ↗ · 2023-12-14 08:00 · UTC+8
Weak-to-strong generalization
公司动态 OpenAI 原文语言:英文
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?
原文摘录
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?
以下是来源提供的简短摘录,完整内容请查看原文。