Creative Priming separates the creative process from the response
Original title: Creative Priming: Decoupling Process from Product to Escape the Creative Paradox
A GRPO-trained priming model writes prompt-specific primers so a response model yields novel, high-quality, diverse outputs.
The paper argues that learning from past creative examples makes those examples recur, which works against creativity. It treats creativity as a process and introduces creative priming: a priming model generates a prompt-specific primer for a separate response model. The primer is trained with GRPO, rewarding primers that produce novel, high-quality, and diverse responses rather than reinforcing any single output.
- The authors call repeating past creative successes the creative paradox.
- Creative priming explicitly separates process learning from response generation.
- GRPO rewards primers that lead to novel, high-quality, and diverse responses.
From the original article
1 min read at huggingface.co
Many of the most consequential advances promised by AI, from curing diseases and inventing new technologies to formulating new scientific theories, are fundamentally creative acts. Yet creativity is uniquely difficult to learn because it faces a creative paradox: learning to repeat past examples of creative success makes those specific examples more likely to recur, inhibiting the very creativity desired.
By Alon Albalak, Sebastian Gabriel, David B. D'Ambrosio · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.