Recently, ChatGPT has attracted great attention, as it can generate fluent and high-quality responses to human inquiries. Several prior studies have shown that ChatGPT attains remarkable generation ability compared with existing models. However, the quantitative analysis of ChatGPT's understanding ability has been given little attention. In this report, we explore the understanding ability of ChatGPT by evaluating it on the most popular GLUE benchmark, and comparing it with 4 representative fine-tuned BERT-style models. We find that: 1) ChatGPT falls short in handling paraphrase and similarity tasks; 2) ChatGPT outperforms all BERT models on inference tasks by a large margin; 3) ChatGPT achieves comparable performance compared with BERT on sentiment analysis and question answering tasks. Additionally, several bad cases from inference tasks show the potential limitation of ChatGPT.

翻译：近日，ChatGPT因其能生成流畅且高质量的人类对话回复而备受关注。多项前期研究表明，相较于现有模型，ChatGPT展现出卓越的生成能力。然而，针对ChatGPT理解能力的量化分析却鲜有关注。本报告通过评估ChatGPT在主流GLUE基准测试上的表现，并将其与四种代表性微调BERT风格模型进行对比，深入探讨其理解能力。研究发现：1）ChatGPT在释义与相似度任务中表现欠佳；2）在推理任务中，ChatGPT以显著优势超越所有BERT模型；3）在情感分析与问答任务中，ChatGPT与BERT表现相当。此外，来自推理任务的若干反面案例揭示了ChatGPT可能存在的局限性。