Hacker News new | ask | show | jobs
by alexarena 550 days ago
This is very cool. It seems like the prompt is asking the LLM to one shot an answer. Have you tried asking it to make a group, confirm whether it's correct, and repeat with the remaining words? (like a human would)