Hacker News new | ask | show | jobs
by andrewgazelka 1115 days ago
My understanding was training on ChatGPT output was against OpenAI ToS. Is this incorrect for this use case (training BERT)?