DarkIntern

[minimum-generated-pig]

Project Overview

This project is based on the InternLM2.5 series model and aims to create a generative dialogue system with a dark, rebellious style. By combining Retrieval-Augmented Generation (RAG) technology, we have built a model capable of generating dialogue filled with villainous traits, such as threats, provocation, and coldness. The system retrieves relevant text information from a pre-constructed knowledge base, enriching its outputs with variety and depth.

Key Features

• Dark Style Dialogue Generation: The model generates responses with a dark, villainous tone, suitable for applications in games, movies, novels, etc.
• RAG-Enhanced Generation: By incorporating external knowledge sources, the model's responses are more diverse and insightful.
• Customizable Dialogue Style: This model is particularly well-suited for users interested in villainous characters, dark philosophy, historical figures, or similar themes.