Content Reconstruction: The Evolution of Texts through Semantic Networks and LLMs

Abstract

This paper presents a novel methodology for generating texts with new meanings through the modification of semantic networks constructed using large language models (LLMs). The approach involves building a semantic network from an initial text, modifying this network by adding and removing nodes and connections, and then generating new texts based on the modified network. The paper outlines the mathematical model, application scenarios, prompts for implementation, and methods for evaluating the generated texts.

Keywords: Large language models, Text generation, AI content creation, Semantic networking, Text reconstruction