Dr. Seg: Revisiting GRPO Training for Visual Large Language Models through Perception-Oriented Design Paper • 2603.00152 • Published Feb 25 • 2
V-Zero: Answer-Label-Free On-Policy Distillation with Contrastive Evidence Gating for Fine-Grained Visual Reasoning Paper • 2606.25319 • Published 17 days ago • 27
V-Zero: Answer-Label-Free On-Policy Distillation with Contrastive Evidence Gating for Fine-Grained Visual Reasoning Paper • 2606.25319 • Published 17 days ago • 27
