RealityBench: Simulated Environments to Evaluate a Vision Language Model's Ability to Document Human Physical Work

cs.CV, cs.AI, cs.CL

Submitted: 2026-10-09

Updated: 2026-10-14

Comments: 28 pages, 3 figures, 14 tables. Data: https://huggingface.co/datasets/relantic/realitybench

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Sources

• Are Large Pre-trained Vision Language Models Effective Construction Safety Inspectors?
• OCRBench v2: An Improved Benchmark for Evaluating Large Multimodal Models on Visual Text Localization and Reasoning
• An Egocentric Vision-Language Model based Portable Real-time Smart Assistant

Related papers

• Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
• AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
• Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
• MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
• TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
• A Survey on Efficient Vision-Language-Action Models