Separating Capability from Confidence: Grounded Dual-State Calibration for GRPO-Trained Medical Vision-Language Models
DualRead builds on the insight that reliability can be read from the actor's internal states at critical moments in the answering process and improves correctness discrimination and calibration over verbalized confidence while preserving answer accuracy.
Yang-Yang Xie, Ke-Chun Hao, Jia-Qi Liu et al.
· 0 citations