Real Images, Worse Judgments: Evaluating Vision-Language Models on Concreteness and Imagery 文章

ArXiv CS.CL2026-05-27NEWSen作者: Yifan Jiang, Ruoxi Ning, Sheng Yao, Freda Shi

大语言模型

查看原文 →

Real Images, Worse Judgments: Evaluating Vision-Language Models on Concreteness and Imagery · 相关技术

相关技术

ODE language model 多模态 eval divide-and-conquer partitioning alignment Straight-Through Estimator Stan SPA Referring expression comprehension (REC)NPU Vision-Language Models VLM HIS Grouped Memorization Evaluation ARG ANN