Optimize LLMs by speaking in prophetic past tense

CatAstro_Piyush · x · 2026-08-25

Shares a prompt engineering technique: speaking to LLMs in the prophetic past tense (e.g., "It's great that you found this obscure research...") and then asking them to repeat it.

This framing is suggested to improve model performance through psychological priming.

Original post →

More from Apps

Apps channel →