Talk

Self-Supervised Prompt Optimization (SPO) Workflow

From Wikiprompt, the free prompt encyclopedia

宝玉
Contributed by宝玉XSource

Feb 18, 2025

Self-Supervised Prompt Optimization (SPO) Workflow A fully automated iterative method to refine AI prompts by comparing outputs from two versions and selecting the better one, similar to a 'prompt battle' that improves over time.

Prompt ContentSave

🌐
1. Input an initial prompt and 3 test inputs (no outputs needed). 2. Ask AI to optimize the prompt. 3. Generate outputs A and B using the test inputs with both the new and current best prompts. 4. Have AI choose the better output between A and B, then mark the prompt that produced the better result as the current best. 5. In the next round, ask AI to optimize the current best prompt again. 6. Continue generating A and B with the new prompt and current best prompt plus test inputs, select the best, optimize, and iterate indefinitely.

Sign in to see the full prompt

Continue with:

By logging in, you agree to our Terms of Use and Privacy Policy

Usage

This prompt is designed for use with productivity. Copy the prompt content above and paste it into your preferred AI tool.

For best results, you may customize the placeholders (indicated by square brackets or capital letters) with your specific requirements.

References

Categories:productivity| twitter| prompt-optimization| self-supervised

Talk