Learning to summarize with human feedback 文章

We’ve applied reinforcement learning from human feedback to train language models that are better at summarization.