Stabilizing RLHF through Advantage Model and Selective Rehearsal.

Baolin Peng Linfeng Song Ye Tian Lifeng Jin Haitao Mi Dong Yu

Published in: CoRR (2023)

Keyphrases

computational model
probabilistic model
theoretical analysis
management system
probability distribution
agent model
hybrid model
bayesian framework
closed form
parameter estimation
real time
cost function
prior knowledge
objective function
high level
information systems
data sets