How AI Learns from Human Preferences
by Writings, Papers and Blogs on Text ModelsAugust 25th, 2024
Too Long; Didn't Read
RLHF is a method to train AI models using human feedback. It involves three phases: 1) training on labeled data, 2) gathering human preferences, and 3) optimizing the model based on those preferences. The Bradley-Terry model is often used to analyze human preferences.
Authors:
(1) Rafael Rafailo, Stanford University and Equal contribution; more junior authors listed earlier;
(2) Archit Sharma, Stanford University and Equal contribution; more junior authors listed earlier;
(3) Eric Mitchel, Stanford University and Equal contribution; more junior authors listed earlier;
(4) Stefano Ermon, CZ Biohub;
(5) Christopher D. Manning, Stanford University;
(6) Chelsea Finn, Stanford University.
Table of Links
4 Direct Preference Optimization
7 Discussion, Acknowledgements, and References
A Mathematical Derivations
A.1 Deriving the Optimum of the KL-Constrained Reward Maximization Objective
A.2 Deriving the DPO Objective Under the Bradley-Terry Model
A.3 Deriving the DPO Objective Under the Plackett-Luce Model
A.4 Deriving the Gradient of the DPO Objective and A.5 Proof of Lemma 1 and 2
B DPO Implementation Details and Hyperparameters
C Further Details on the Experimental Set-Up and C.1 IMDb Sentiment Experiment and Baseline Details
C.2 GPT-4 prompts for computing summarization and dialogue win rates
D Additional Empirical Results
D.1 Performance of Best of N baseline for Various N and D.2 Sample Responses and GPT-4 Judgments
3 Preliminaries
We review the RLHF pipeline in Ziegler et al. (and later [38, 1, 26]). It usually includes three phases: 1) supervised fine-tuning (SFT); 2) preference sampling and reward learning and 3) RL optimization.
This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license.
L O A D I N G
. . . comments & more!
. . . comments & more!
