We feed the extracted evidence sentences into exist- ing MRC models and evaluate the end-to-end performance on three challenging multiple- choice MRC datasets: ...
確定! 回上一頁