MLA-Trust Dataset: GUI Environment Data for Multimodal LLM Agent Trustworthiness Evaluation
收藏IEEE2026-04-17 收录
数据链接:
官方服务:
资源简介:
This paper presents a comprehensive taxonomy for evaluating the trustworthiness of multimodal large language model (LLM) agents operating in graphical user interface (GUI) environments. Our framework identifies four primary dimensions of trustworthiness: Controllability, Privacy, Safe Data, and Truthfulness.




