3 ms·Post-Training Generative Recommenders with Advantage-Weighted Supervised Tuning1 points by CharlesW 1y ago