#91 · Primary category: AI Design & Prototyping

dreamGPT

agents artificial-intelligence automation generative-model ideas llm llms machinelearning ml

Leverage hallucinations from Large Language Models (LLMs) for novelty-driven explorations.

Project last updated:05/07/24

GitHub Stars

591

Forks

39

Contributors

5

License

MIT

Why we included this project

dreamGPT treats LLM hallucinations as a feature rather than a bug. Instead of answering a specific question, it amplifies the hallucinations most users try to suppress, and uses them to generate new ideas. You can pass an optional theme, and the tool starts from a random set of concepts, writes titled idea descriptions, scores each one on several criteria, and feeds the strongest results back into the next round. The result is an iterative brainstorming loop that can surface ideas you would not have reached on your own. It's a lightweight Python tool that needs an OpenAI API key and optionally Pinecone for storing dreams, so it's most useful to developers and researchers who want to prototype a divergent-thinking pipeline or study how to steer LLM creativity.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category