• Self-Refine: Iterative Refinement with Self-Feedback - a novel approach that allows LLMs to iteratively refine outputs and incorporate feedback along multiple dimensions to improve performance on diverse tasks https://arxiv.org/pdf/2303.17651.pdf 5 comments 3/4/2023 machinelearning