Dataset for "Interpreting Neural CWI Classifiers' Weights as Vocabulary Size." Yo Ehara, BEA 2020.

The dataset is in the CSV format.
The first line shows field names.
From the second line, each line denotes a learner.
The fields starting from "p##" are responses to vocabulary questions.
The fields starting from "l##" and "s##" are responses of reading comprehension questions.
##s denote numbers.

TOEICy, TOEICm are the year and month of the last TOEIC test that the learner took.
TOEICscore is the total TOEIC score of the last TOEIC test.
TOEICl and TOEICr are the listening and reading comprehension TOEIC test scores, respectively. Filling these fields was optional for this dataset.

This attachment does not include the problems and correct options for each question due to copyright reasons.
The information on how to obtain these, please refer to:
http://yoehara.com/vocabulary-prediction/