This repo contains various code and datasets used for the two probing studies discussed below.
Repository for semantic fragment experiments in the following paper: Probing Natural Language Inference Models through Semantic Fragments[https://arxiv.org/pdf/1909.07521.pdf] (AAAI 2020)
For reproducing the experiments, please see the README in
scripts
. More information about the different datasets see data
and the README there.
Main Contributors Hai Hu (huhai@indiana.edu), Larry Moss (lmoss@indiana.edu),Kyle Richardson (kyler@allenai.org), Ashish Sarharwal (ashishs@allenai.org)
This repo contains the data and code for the paper What Does My QA Model Know: DEvising Controlled Probes using Expert Knowledge[https://arxiv.org/abs/1912.13337] (TACL 2020)
For reproducing the experiments, please see README in scripts_mcqa
(with pointers to original experimental data), and the raw datasets in data_mcqa
.
Main Contributors Kyle Richardson (kyler@allenai.org), Ashish Sarharwal (ashishs@allenai.org)