Evaluation of natural language generation (NLG) is complex and multi-dimensional. Generated text can be evaluated for fluency, coherence, factuality, or any other dimensions of interest. Most frameworks that perform such multi-dimensional evaluation require training on large manually or synthetically generated datasets. In this paper, we study the efficacy of large language models as multi-dimensional evaluators using in-context learning, obviating the need for large training datasets. Our experiments show that in-context learning-based evaluators are competitive with learned evaluation frameworks for the task of text summarization, establishing state-of-the-art on dimensions such as relevance and factual consistency. We then analyze the effects of factors such as the selection and number of in-context examples on performance. Finally, we study the efficacy of in-context learning based evaluators in evaluating zero-shot summaries written by large language models such as GPT-3.
@article{arxiv.2306.01200,
title = {Multi-Dimensional Evaluation of Text Summarization with In-Context Learning},
author = {Sameer Jain and Vaishakh Keshava and Swarnashree Mysore Sathyendra and Patrick Fernandes and Pengfei Liu and Graham Neubig and Chunting Zhou},
journal= {arXiv preprint arXiv:2306.01200},
year = {2024}
}