LLM feedback framework for prompt optimization, rather self improving agent and back propagation with text gradient is just a bad analogy.
arxiv.org
https://arxiv.org/pdf/2406.07496
TextGrad
Creator
Creator
Seonglae ChoCreated
Created
2025 Jun 13 18:55Editor
Editor
Seonglae ChoEdited
Edited
2025 Jun 13 19:1Refs
Refs