Policy constraints are the critical factor for stable PPO training in RLHF, and the proposed PPO-max variant improves stability for large language model alignment.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2023 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Secrets of RLHF in Large Language Models Part I: PPO
Policy constraints are the critical factor for stable PPO training in RLHF, and the proposed PPO-max variant improves stability for large language model alignment.