1 paper · 1 filter
Buxin Su, Bingxuan Li, Cheng Qian +3
Supervised fine-tuning with synthetic rationale data is widely assumed to improve language model performance on clinical prediction tasks by teaching models not just what to predic…