3 ms·
Wouldn't it be easier to train a second model to check whether the answer is acceptable, and declining to answer if it isn't?
by joelthelion 3y ago
Wouldn't it be easier to train a second model to check whether the answer is acceptable, and declining to answer if it isn't?