Ji-Kai Wang

Criterion-Conditional In-Context Learning: Evaluating Criterion-Shift Adaptation in Vision-Language Models

This work introduces a new setting, denoted as Criterion-Conditional In-Context Learning (CC-ICL), where models must infer the latent criterion from context and adjust predictions accordingly under fixed task semantics, and proposes two complementary metrics, Criterion Invariance and Criterion Sensitivity, to evaluate...