Multi-Reward RL, Part 3: GDPO + CISPO + REPO-R at 27B, and the Advantage Floor That Stopped Learning

Part 1 explained how PPO, GRPO, DAPO and GDPO turn several rewards into one learning signal. Part 2 benchmarked the trainers on a 14B model for 50 steps and found a trap: CISPO won the training curves