Can Large Language Model Summarizers Adapt to Diverse Scientific Communication Goals?

arXiv:2401.10415v2 Announce Type: replace
Abstract: In this work, we investigate the controllability of large language models (LLMs) on scientific summarization tasks. We identify key stylistic and content coverage factors that characterize different types of summaries such as paper reviews, abstracts, and lay summaries. By controlling stylistic features, we find that non-fine-tuned LLMs outperform humans in the MuP review generation task, both in terms of similarity to reference summaries and human preferences. Also, we show that we can improve the controllability …

abstract adapt arxiv communication coverage cs.ai cs.cl diverse features identify key language language model language models large language large language model large language models llms paper replace reviews scientific summarization tasks type types work