一作ACL 2026Poster· Natural Language Processing
A uto J udger: An Agent-Driven Framework for Efficient Benchmarking of MLLM s
Xuanwen Ding · Chengjun Pan · Zejun Li · Jiwen Zhang · Siyuan Wang · zhongyu wei
摘要
Evaluating multimodal large language models (MLLMs) is becoming increasingly expensive as benchmarks grow in scale and cross-modality complexity. Inspired by structuralism in cognitive psychology, we tackle this difficulty with an adaptive evaluation framework for efficient benchmarking, namely A...