04:00
2026-09-02
arxiv.org
artificial-intelligence
ReNFT: Repairing Mode Collapse in Reward Post-Training via Internal Probability-Mass Recalibration
Researchers propose ReNFT, a method that repairs mode collapse in reward post-training of diffusion generators by recalibrating internal probability mass, preserving 98.9% and 99.0% of NFT's reward onβ¦