SOURCE-LINKED INTELLIGENCE

GroundBench: A Factorized, Counterfactual Benchmark for Locating VLM Affordance Failures

arXiv · AI, language, vision and robotics · article · Sep 10, 2026 · UTC

A companion evaluation found that naming the target part in a manipulation prompt increased action accuracy by 0.32-0.63 across eight vision-language models, with no model outperforming a constant baseline until the part was named. However, naming the part supplies information that a real system must infer, confounding visual grounding, mechanical reasoning, and category-to-action association. We introduce GroundBench, a diagnostic benchmark that separates these explanations through six branch-and-merge conditions, each adding a controlled information bundle, and a counterfactual re-ask target

Read original source ↗ Open in workspace

recordType

paper

region

Global