English
Share
Sign In
React: Techniques for adding inference and action
ReAct Prompt is a new framework that combines the inference and behavior of language models to improve them. This approach allows language models to not only generate thought processes, but also interact with external sources to obtain additional information and derive reliable responses based on this.
"뉴진스 맴버는 누가 있나요?"How it works
1.
LM first infers basic information about the news.
2.
Based on the inference, we retrieve information about the members of NewJeans from external databases (e.g. Wikipedia).
3.
Integrate search results with internal inference to generate accurate, detailed answers to your questions.
The members of New Jeans are Minzy, Hani, Daniel, Haerin, and Hyein.
ReAct prompts are more effective for knowledge-intensive tasks than traditional Chain-of-Thought (CoT) methods by integrating external information, and can increase human interpretation and trust in the problem-solving process.
ReAct Prompting is a new framework that allows language models to go beyond simple text-based inference and interact with external sources to obtain additional information and incorporate it into the inference process. This expands the capabilities of language models and allows them to provide more reliable and realistic answers. ReAct Prompting plays a key role in improving the adaptability and accuracy of models, especially in knowledge-intensive or complex tasks.
ⓒ 2023. Haebom, all rights reserved.
It may be used for commercial purposes with permission from the copyright holder, provided the source is cited.
