Comment by smokel

You must be confused. This research is about reinforcement learning, not about large language models.