sp26-cs639 cs639
Lecture 17: Alignment I Lecture 17: Alignment I
Content locked
The catalog is public. Reading this content requires approved access.
Sign in to readLecture 17: Alignment I
From adaptation to alignment
The RLHF pipeline
1. Collect comparisons
2. Fit a scalar reward model
3. Optimize the language model
Reinforcement-learning refresher
Why might preference-based training help?
Connections
Source figure