Can LLMs correct themselves? a benchmark of self-correction in LLMs

Published in NeurIPS, 2025