Viewing a single comment thread. View all comments

LastVariation t1_jdsx10i wrote

Start by training a gpt-2, then add 2, then reinforce it to not be crazy by feeding chatGPT responses

3