Hacker News new | ask | show | jobs
by drsim 357 days ago
Right. I’ve had a lot of success using structured output to force LLMs to make Boolean choices, like can they reply or not.