Retrieval-augmented generation models for predicting pathological upgrading in endoscopically resected gastric low-grade intraepithelial neoplasia: A dual-center evaluation

Figure 3 Model calibration curves. The dotted line (apparent) represents the performance of the retrieval-augmented generation-enhanced large language models, whereas the solid line (bias-corrected) accounts for and adjusts potential biases. The diagonal reference line (ideal) indicates perfect calibration, where the predictions match the observed outcomes exactly. GLM: Generalized linear mixed.