1 paper
Elroy Stav, Dvir Berlowitz, Maayan Orner +1
Intrinsic self-correction (SC) aims to improve large language model outputs by prompting a model to revisit its own initial answer without external feedback. Recent studies have qu…