ElitePLM: An Empirical Study on General Language Ability Evaluation of Pretrained Language Models

TL;DR

This paper conducts a comprehensive empirical evaluation of ten popular pretrained language models across four key language ability dimensions, providing insights into their strengths, limitations, and transferability for NLP tasks.

Contribution

It introduces a large-scale, systematic evaluation framework for assessing the general language abilities of various PLMs across multiple dimensions.

Findings

PLMs excel in different ability tests based on their training objectives.

Fine-tuning sensitivity varies with data size and distribution.

PLMs show strong transferability between similar tasks.

Abstract

Nowadays, pretrained language models (PLMs) have dominated the majority of NLP tasks. While, little research has been conducted on systematically evaluating the language abilities of PLMs. In this paper, we present a large-scale empirical study on general language ability evaluation of PLMs (ElitePLM). In our study, we design four evaluation dimensions, i.e. memory, comprehension, reasoning, and composition, to measure ten widely-used PLMs within five categories. Our empirical results demonstrate that: (1) PLMs with varying training objectives and strategies are good at different ability tests; (2) fine-tuning PLMs in downstream tasks is usually sensitive to the data size and distribution; (3) PLMs have excellent transferability between similar tasks. Moreover, the prediction results of PLMs in our experiments are released as an open resource for more deep and detailed analysis on the language abilities of PLMs. This paper can guide the future work to select, apply, and design PLMs for specific tasks. We have made all the details of experiments publicly available at https://github.com/RUCAIBox/ElitePLM.