Edit, Don't Imitate: Minimal-Edit On-Policy Distillation for Multimodal Reasoning.
arXiv:2610.30231v1 Announce Type: cross Abstract: Visual mathematical reasoning can fail before reasoning begins: when a vision-language model (VLM) misreads a diagram, subsequent reasoning proceeds from false premises. This perceptual bottleneck is especially
arXiv cs.AIResearch source·