Hacker News new | ask | show | jobs
by FartyMcFarter 2059 days ago
Wouldn't this kind of priming be brittle and unreliable? Has anyone successfully primed GPT-3 to solve complex problems consistently?