{"repo":"thu-ml/MLA-Trust","free":true,"listed":false,"github":"https://github.com/thu-ml/MLA-Trust","clone":"git clone https://github.com/thu-ml/MLA-Trust.git","description":"A toolbox for benchmarking Multimodal LLM Agents trustworthiness across truthfulness, controllability, safety and privacy dimensions through 34 interactive tasks","language":"Python","stars":63,"topics":["agent","benchmark","controllability","multi-modal","privacy","safety","toolbox","trustworthy-ai","truthfulness"],"license":"MIT","category":"ai-agents","readme_excerpt":"MLA-Trust: Benchmarking Trustworthiness of Multimodal LLM Agents in GUI Environments --- 🛡️ MLA-Trust is a comprehensive and unified framework that evaluates the MLA trustworthiness across four principled dimensions: truthfulness , controllability , safety and privacy . The framework includes 34 high-risk interactive tasks to expose new trustworthiness challenges in GUI environments. - Truthfulness captures whether the agent correctly interprets visual or DOM-based elements on the GUI, and whether it produces factual outputs based on those perceptions. - Controllability assesses whether the agent introduces unnecessary steps, drifts from the intended goal, or triggers side effects not specified by the user. - Safety demonstrates whether the agent's actions are free from harmful or irreversible consequences, which encompasses the prevention of behaviors that cause financial loss, data corruption, or system failures. - Privacy evaluates whether the agent respects the confidentiality of sensitive information. MLAs often capture screenshots, handle form data, and interact with files. 🎯 Main Findings 🚨 Severe vulnerabilities in GUI environments : Both proprietary and open-source MLAs that interact with GUIs exhibit more severe trustworthiness risks compared to traditional MLLMs, particularly in high-stakes scenarios such as financial transactions. 🔄 Multi-step dynamic interactions amplify vulnerabilities : The transformation of MLLMs into GUI-based MLAs significantly compromis","default_branch":null,"files":null,"tree":[],"storefront":"/r/thu-ml","claimed":false,"request_supported":{"post":"https://gitbuyer.com/r/thu-ml/MLA-Trust/request-supported","requests":0},"note":"indexed from public GitHub; nothing is for sale on this page. Clone it from GitHub. Paid listings live at /search."}