# CVT-Bench: Probing Spatial-State Integrity through Counterfactual Viewpoint Transformations

arXiv: https://arxiv.org/abs/2603.21114 · HTML: https://mlsanity.com/paper/2603.21114/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Embodied and 3D vision-language models, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 20 of its 27 references that are in the map.
- First posted on arXiv 22 Mar 2026.

## Placed in

- Multimodal LLMs / vision-language models / [Embodied and 3D vision-language models](https://mlsanity.com/mllm/embodied-and-3d-vision-language-models/)

## Builds on (most cited first, 20 of 20)

- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [Lost in the Middle: How Language Models Use Long Contexts](https://mlsanity.com/paper/2307.03172/) (2023)
- [VQA: Visual Question Answering](https://mlsanity.com/paper/1505.00468/) (2015)
- [CLEVR: A Diagnostic Dataset for Compositional Language and Elementary Visual Reasoning](https://mlsanity.com/paper/1612.06890/) (2016)
- [LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding](https://mlsanity.com/paper/2308.14508/) (2023)
- [RULER: What's the Real Context Size of Your Long-Context Language Models?](https://mlsanity.com/paper/2404.06654/) (2024)
- [Eyes Wide Shut? Exploring the Visual Shortcomings of Multimodal LLMs](https://mlsanity.com/paper/2401.06209/) (2024)
- [SpatialVLM: Endowing Vision-Language Models with Spatial Reasoning Capabilities](https://mlsanity.com/paper/2401.12168/) (2024)
- [Winoground: Probing Vision and Language Models for Visio-Linguistic Compositionality](https://mlsanity.com/paper/2204.03162/) (2022)
- [L-Eval: Instituting Standardized Evaluation for Long Context Language Models](https://mlsanity.com/paper/2307.11088/) (2023)
- [Learning 3D Semantic Scene Graphs from 3D Indoor Reconstructions](https://mlsanity.com/paper/2004.03967/) (2020)
- [MuirBench: A Comprehensive Benchmark for Robust Multi-image Understanding](https://mlsanity.com/paper/2406.09411/) (2024)
- [MindCube: Spatial Mental Modeling from Limited Views](https://mlsanity.com/paper/2506.21458/) (2025)
- [LLaVA-ST: A Multimodal Large Language Model for Fine-Grained Spatial-Temporal Understanding](https://mlsanity.com/paper/2501.08282/) (2025)
- [MC-Bench: A Benchmark for Multi-Context Visual Grounding in the Era of MLLMs](https://mlsanity.com/paper/2410.12332/) (2024)
- [SpatialBench: Benchmarking Multimodal Large Language Models for Spatial Cognition](https://mlsanity.com/paper/2511.21471/) (2025)
- [SpinBench: Perspective and Rotation as a Lens on Spatial Reasoning in VLMs](https://mlsanity.com/paper/2509.25390/) (2025)
- [Struct2D: A Perception-Guided Framework for Spatial Reasoning in MLLMs](https://mlsanity.com/paper/2506.04220/) (2025)
- [AQA-Bench: An Interactive Benchmark for Evaluating LLMs' Sequential Reasoning Ability](https://mlsanity.com/paper/2402.09404/) (2024)
- [LRR-Bench: Left, Right or Rotate? Vision-Language models Still Struggle With Spatial Understanding Tasks](https://mlsanity.com/paper/2507.20174/) (2025)

## Built on by (most cited first, 2 of 2)

- [Is Your Trajectory Displacement Safe in Long-tail?](https://mlsanity.com/paper/2606.16313/) (2026)
- [CrossProjection: Geometric Grounding Beyond Viewpoint Change in Architectural Drawings](https://mlsanity.com/paper/2608.00473/) (2026)

## Abstract (arXiv)

Multimodal large language models (MLLMs) perform strongly on isolated spatial tasks, but whether their predictions remain persistent and mutually coherent across viewpoints and competing scenes is unclear. We formalize this behavioral property as spatial-state integrity and introduce CVT-Bench, a factorial diagnostic suite spanning two domains (CVT-Synthetic, CVT-Real), four context regimes (CVT-Isolated, irrelevant filler, attribute filler, CVT-Competing), three representations (Image, Text/BBox, Scene Graph), and ten viewpoint conditions (nine azimuths from 0^circ to 360^circ at 45^circ increments, plus Top) with target views withheld. Across 200 scenes and 11,833 relational queries, we measure counterfactual accuracy, cycle consistency, the normalized survival coefficient, spatial realizability, and context-specific interference. Five state-of-the-art MLLMs exhibit rapid persistence loss and frequently produce jointly unrealizable states despite high local accuracy, with failures amplified by competing contexts and natural-scene complexity. Matched irrelevant and attribute filler controls show that prompt length, position, or loss of scene access alone are insufficient explanations. On average, Text/BBox improves accuracy, persistence, and realizability, while Scene Graph provides complementary gains; neither eliminates the instability. Thus, isolated spatial accuracy substantially overestimates robustness, establishing spatial-state integrity as a distinct evaluation target. The complete benchmark and codebase will be publicly released.