This post discusses the use of reflection in AI systems, such as LLMs. It explores different techniques, including basic reflection, reflexion, and Language Agent Tree Search, that can improve the quality of outputs and responses. The post emphasizes the importance of output quality over speed and explains how these techniques can be applied.

5m read time From blog.langchain.dev
Post cover image
Table of contents
Basic ReflectionReflexionLanguage Agent Tree SearchConclusion

Sort: