Reinforcement Fine-Tuning

Researchers are exploring its potential to improve alignment in large language models.