Hacker News new | ask | show | jobs
by popinman322 507 days ago
Assuming you're doing local inference, have you tried setting a token filter on the model?