Skip navigation
Title: 

BLiMP: The Benchmark of Linguistic Minimal Pairs for English (Electronic Resources)

Authors: Warstadt, Samuel R.
Parrish, Alicia
Liu, Haokun
Mohananey, Anhad
Peng, Wei
Wang, Sheng-Fu
Bowman, Samuel R.
Keywords: computational linguistics
Issue Date: Jul-2020
Publisher: The MIT Press
Citation: BLiMP: The Benchmark of Linguistic Minimal Pairs for English Alex Warstadt, Alicia Parrish, Haokun Liu, Anhad Mohananey, Wei Peng, Sheng-Fu Wang, and Samuel R. Bowman Transactions of the Association for Computational Linguistics 2020 Vol. 8, 377-392
Abstract: We introduce The Benchmark of Linguistic Minimal Pairs (BLiMP),1 a challenge set for evaluating the linguistic knowledge of language models (LMs) on major grammatical phenomena in English. BLiMP consists of 67 individual datasets, each containing 1,000 minimal pairs—that is, pairs of minimally different sentences that contrast in grammatical acceptability and isolate specific phenomenon in syntax, morphology, or semantics. We generate the data according to linguist-crafted grammar templates, and human aggregate agreement with the labels is 96.4%. We evaluate n-gram, LSTM, and Transformer (GPT-2 and Transformer-XL) LMs by observing whether they assign a higher probability to the acceptable sentence in each minimal pair. We find that state-of-the-art models identify morphological contrasts related to agreement reliably, but they struggle with some subtle semantic and syntactic phenomena, such as negative polarity items and extraction islands.
URI: http://hdl.handle.net/2451/61422
Appears in Collections:Machine Learning for Language Lab

Files in This Item:
File Description SizeFormat 
blimp-master.zipBLiMP dataset and paper preprint6.6 MBUnknownView/Open
blimp_ngram-master.zipn-Gram Model1.73 kBUnknownView/Open
colorlessgreenRNNs-master.zipLSTM Model25.37 MBUnknownView/Open
data_generation-master.zipBLiMP data generation code4.82 MBUnknownView/Open
jiant-blimp-and-npi.zipGPT-2 and Transformer XL models (see scripts/blimp subdirectory)593.36 kBUnknownView/Open


Items in FDA are protected by copyright, with all rights reserved, unless otherwise indicated.