4 ms·
"models primed with incorrect solutions containing proper reasoning patterns achieve comparable performance to those trained on correct solutions" One of the p
by nickpsecurity 2y ago
"models primed with incorrect solutions containing proper reasoning patterns achieve comparable performance to those trained on correct solutions"
One of the parts most worth a replication study.