How Can We Know What Language Models Know?

Zhengbao Jiang, Frank F. Xu, Jun Araki, Graham Neubig


Abstract
Recent work has presented intriguing results examining the knowledge contained in language models (LMs) by having the LM fill in the blanks of prompts such as “Obama is a __ by profession”. These prompts are usually manually created, and quite possibly sub-optimal; another prompt such as “Obama worked as a __ ” may result in more accurately predicting the correct profession. Because of this, given an inappropriate prompt, we might fail to retrieve facts that the LM does know, and thus any given prompt only provides a lower bound estimate of the knowledge contained in an LM. In this paper, we attempt to more accurately estimate the knowledge contained in LMs by automatically discovering better prompts to use in this querying process. Specifically, we propose mining-based and paraphrasing-based methods to automatically generate high-quality and diverse prompts, as well as ensemble methods to combine answers from different prompts. Extensive experiments on the LAMA benchmark for extracting relational knowledge from LMs demonstrate that our methods can improve accuracy from 31.1% to 39.6%, providing a tighter lower bound on what LMs know. We have released the code and the resulting LM Prompt And Query Archive (LPAQA) at https://github.com/jzbjyb/LPAQA.
Anthology ID:
2020.tacl-1.28
Volume:
Transactions of the Association for Computational Linguistics, Volume 8
Month:
Year:
2020
Address:
Cambridge, MA
Editors:
Mark Johnson, Brian Roark, Ani Nenkova
Venue:
TACL
SIG:
Publisher:
MIT Press
Note:
Pages:
423–438
Language:
URL:
https://aclanthology.org/2020.tacl-1.28
DOI:
10.1162/tacl_a_00324
Bibkey:
Cite (ACL):
Zhengbao Jiang, Frank F. Xu, Jun Araki, and Graham Neubig. 2020. How Can We Know What Language Models Know?. Transactions of the Association for Computational Linguistics, 8:423–438.
Cite (Informal):
How Can We Know What Language Models Know? (Jiang et al., TACL 2020)
Copy Citation:
PDF:
https://aclanthology.org/2020.tacl-1.28.pdf
Code
 jzbjyb/LPAQA
Data
LAMA