In our attack, we insert adjectives and adverbs into natural places of incorrect student answers, fooling the model into predicting them as ...
確定! 回上一頁