Newest papers

VERA: A Responsible AI Dashboard for Assessing LLM Deployments in Light of the EU AI Act

Robin Quériaux, Solveig Bachellery, Léa Déléris, Saurabhsingh Rajput, Mariam Barry, Tushar Sharma, Jérôme Lebecq · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.31588 · cited by 0 works (OpenAlex) · open-access PDF

Spanish

Las instituciones que desplegan grandes modelos de lenguaje en entornos regulados a menudo necesitan pruebas técnicas de prácticas de IA responsables antes de aprobar un despliegue. COMPL-AI proporciona una interpretación técnica de la Ley de IA de la UE que organiza esta evidencia bajo seis grupos de principios y dieciocho requisitos técnicos, de los cuales doce pueden medir los benchmarks actuales. Las suites de benchmarks que producen esta evidencia son típicamente operadas desde una línea de comandos, mientras que el personal que aprueba los benchmarks son no ingenieros públicos. Presentamos VERA, un panel de control de código abierto autogestionado desplegado para facilitar los procesos de IA responsables dentro de un gran banco europeo. Un revisor selecciona un modelo y ejecuta los benchmarks públicos contra él. VERA estudia los doce requisitos

Machine translation (Meta M2M100); the English original follows.

Abstract

Institutions that deploy large language models in regulated settings often need technical evidence of responsible AI practices before approving a deployment. COMPL-AI provides a technical interpretation of the EU AI Act that organises this evidence under six principle groups and eighteen technical requirements, twelve of which current benchmarks can measure. The benchmark suites that produce this evidence are typically operated from a command line, while the staff who approve deployments are non-engineers. We present VERA, an open-source self-hosted dashboard deployed to facilitate responsible AI processes within a large European bank. A reviewer selects a model and runs public benchmarks against it. VERA scores the twelve requirements that current benchmarks can measure and instruments the energy consumed by each evaluation, adding run-level evidence for environmental impact. We run four served open-weight models through the requirement pipeline, and the dashboard displays a score for each measurable requirement. A within-subject study then compares how participants read these benchmark results: from the raw output files in one condition, and through the VERA dashboard in the other. The dashboard reduced the time needed to perform the responsible AI assessment, reducing the time required from 56 seconds using raw files to 30 seconds using our dashboard, on average. Our dashboard scored highly (4.36 out of 5) on usefulness in the survey conducted within the organization. The dashboard and the study material are released publicly to promote replicability and reproducibility in different organisational contexts.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Dashboard, Benchmark (surveying), Raw data, Computer science, Benchmarking, Measure (data warehouse)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange