Automatic Prompt Engineering
A technical direction that uses large models to automatically search for and optimize prompts. It is part of automated model optimization, in which the prompt itself is treated as a variable to be tuned rather than written by hand. Automated search methods explore candidate prompts and select those that improve task performance, reducing reliance on manual prompt design and enabling systematic optimization of model behavior.
Timeline (last 90 days)
In the field of automated model optimization, there are already some comparable real works, such as using large models to search for better prompts, data recipes or training hyperparameters (e.g., automatic prompt engineering, attempts to combine neural architecture search with large models).
What existing automated model optimization works have in common is that the large model outputs suggestions or configurations, and it is still the training pipeline that actually changes the model parameters.
All Facts (2)
In the field of automated model optimization, there are already some comparable real works, such as using large models to search for better prompts, data recipes or training hyperparameters (e.g., automatic prompt engineering, attempts to combine neural architecture search with large models).
50%UnverifiedWhat existing automated model optimization works have in common is that the large model outputs suggestions or configurations, and it is still the training pipeline that actually changes the model parameters.
50%