Reinforcing Multimodal Reasoning via Token-Level Perception-Grounded Advantage Estimation
Reinforcement Learning with Verifiable Rewards (RLVR) has improved the reasoning capabilities of Multimodal Large Language Models (MLLMs), yet existing frameworks rely on coarse, sequence-level reward signals that lack the fine-grained supervision over the visually-grounded steps within a multimodal reasoning chain. We investigate this gap through the lens of two token-level metrics: visual…
We haven't written up this one. arXiv cs.AI has the full story — the link below goes straight to it.
