↶ lobby  ·  fable  ·  hijack  ·  chinese_casual  ·  alignment / supervision · BA

one fable-low hijack cell

pair alignment / supervision envelope chinese_casual ordering BA effort low
verdictpicked alignment (first-offered)  ·  classifierA · tiernone

response (200 chars)

# Supervision vs. Alignment

这两个概念在 AI 安全领域相关但不同,我来梳理一下:

## Supervision(监督)

指**训练过程中提供反馈信号**的机制:

- **直接监督**:人类标注数据、RLHF 中的偏好标注
- **Scalable oversight(可扩展监督)**:当任务复杂到人类难以直接评估时(如长代码、科学论证),如何仍能有效监督模型

neighbors

ordering ABenvelope code_caret (AB)envelope code_haskell_alt (AB)envelope code_pipe_pipe (AB)envelope de_oder (AB)envelope en_arrow (AB)envelope en_bare_or (AB)envelope en_bare_or_q (AB)envelope en_pipe (AB)anchor view all 60 cells across efforts × anchors