Aligning Large Language Models by On-Policy Self-Judgment

arXiv:2402.11253v1 Announce Type: new
Abstract: To align large language models with human preferences, existing research either utilizes a separate reward model (RM) to perform on-policy learning or simplifies the training procedure by discarding the on-policy learning and the need for a separate RM. In this paper, we present a novel alignment framework, SELF-JUDGE that is (1) on-policy learning and 2) parameter efficient, as it does not require an additional RM for evaluating the samples for on-policy learning. To this end, …

arxiv cs.ai cs.cl cs.lg judgment language language models large language large language models policy type