CLiMP: A Benchmark for Chinese Language Model Evaluation

Beilei Xiang, Changbing Yang, Yu Li, Alex Warstadt, Katharina Kann


Abstract
Linguistically informed analyses of language models (LMs) contribute to the understanding and improvement of such models. Here, we introduce the corpus of Chinese linguistic minimal pairs (CLiMP) to investigate what knowledge Chinese LMs acquire. CLiMP consists of sets of 1000 minimal pairs (MPs) for 16 syntactic contrasts in Chinese, covering 9 major Chinese linguistic phenomena. The MPs are semi-automatically generated, and human agreement with the labels in CLiMP is 95.8%. We evaluate 11 different LMs on CLiMP, covering n-grams, LSTMs, and Chinese BERT. We find that classifier–noun agreement and verb complement selection are the phenomena that models generally perform best at. However, models struggle the most with the ba construction, binding, and filler-gap dependencies. Overall, Chinese BERT achieves an 81.8% average accuracy, while the performances of LSTMs and 5-grams are only moderately above chance level.
Anthology ID:
2021.eacl-main.242
Volume:
Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume
Month:
April
Year:
2021
Address:
Online
Editors:
Paola Merlo, Jorg Tiedemann, Reut Tsarfaty
Venue:
EACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
2784–2790
Language:
URL:
https://aclanthology.org/2021.eacl-main.242
DOI:
10.18653/v1/2021.eacl-main.242
Bibkey:
Cite (ACL):
Beilei Xiang, Changbing Yang, Yu Li, Alex Warstadt, and Katharina Kann. 2021. CLiMP: A Benchmark for Chinese Language Model Evaluation. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, pages 2784–2790, Online. Association for Computational Linguistics.
Cite (Informal):
CLiMP: A Benchmark for Chinese Language Model Evaluation (Xiang et al., EACL 2021)
Copy Citation:
PDF:
https://preview.aclanthology.org/ingest-2024-clasp/2021.eacl-main.242.pdf
Dataset:
 2021.eacl-main.242.Dataset.zip
Data
BLiMP