# Why MLLMs Struggle to Determine Object Orientations

arXiv: https://arxiv.org/abs/2604.13321 · HTML: https://mlsanity.com/paper/2604.13321/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in MLLM training and visual instruction tuning, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 32 of its 44 references that are in the map.
- First posted on arXiv 14 Apr 2026.

## Placed in

- Multimodal LLMs / vision-language models / [MLLM training and visual instruction tuning](https://mlsanity.com/mllm/mllm-training-and-visual-instruction-tuning/)

## Builds on (most cited first, 30 of 32)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [ImageNet Large Scale Visual Recognition Challenge](https://mlsanity.com/paper/1409.0575/) (2014)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [Sigmoid Loss for Language Image Pre-Training](https://mlsanity.com/paper/2303.15343/) (2023)
- [LLaVA-OneVision: Easy Visual Task Transfer](https://mlsanity.com/paper/2408.03326/) (2024)
- [InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models](https://mlsanity.com/paper/2504.10479/) (2025)
- [Eyes Wide Shut? Exploring the Visual Shortcomings of Multimodal LLMs](https://mlsanity.com/paper/2401.06209/) (2024)
- [Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall Spaces](https://mlsanity.com/paper/2412.14171/) (2024)
- [SpatialVLA: Exploring Spatial Representations for Visual-Language-Action Model](https://mlsanity.com/paper/2501.15830/) (2025)
- [Spatial-MLLM: Boosting MLLM Capabilities in Visual-based Spatial Intelligence](https://mlsanity.com/paper/2505.23747/) (2025)
- [Visual Interpretability for Deep Learning: a Survey](https://mlsanity.com/paper/1802.00614/) (2018)
- [OmniSpatial: Towards Comprehensive Spatial Reasoning Benchmark for Vision Language Models](https://mlsanity.com/paper/2506.03135/) (2025)
- [Why Is Spatial Reasoning Hard for VLMs? An Attention Mechanism Perspective on Focus Areas](https://mlsanity.com/paper/2503.01773/) (2025)
- [Perception Tokens Enhance Visual Reasoning in Multimodal Language Models](https://mlsanity.com/paper/2412.03548/) (2024)
- [An Empirical Analysis on Spatial Reasoning Capabilities of Large Multimodal Models](https://mlsanity.com/paper/2411.06048/) (2024)
- [DepthVLA: Enhancing Vision-Language-Action Models with Depth-Aware Spatial Reasoning](https://mlsanity.com/paper/2510.13375/) (2025)
- [Embodied-Reasoner: Synergizing Visual Search, Reasoning, and Action for Embodied Interactive Tasks](https://mlsanity.com/paper/2503.21696/) (2025)
- [How Far are VLMs from Visual Spatial Intelligence? A Benchmark-Driven Perspective](https://mlsanity.com/paper/2509.18905/) (2025)
- [Visual Representation Alignment for Multimodal Large Language Models](https://mlsanity.com/paper/2509.07979/) (2025)
- [VisOnlyQA: Large Vision Language Models Still Struggle with Visual Perception of Geometric Information](https://mlsanity.com/paper/2412.00947/) (2024)
- [Multimodal Spatial Reasoning in the Large Model Era: A Survey and Benchmarks](https://mlsanity.com/paper/2510.25760/) (2025)
- [Why Do MLLMs Struggle with Spatial Understanding? A Systematic Analysis from Data to Architecture](https://mlsanity.com/paper/2509.02359/) (2025)
- [Representation Engineering for Large-Language Models: Survey and Research Challenges](https://mlsanity.com/paper/2502.17601/) (2025)
- [Vision-Language Models Can't See the Obvious](https://mlsanity.com/paper/2507.04741/) (2025)
- [Euclid's Gift: Enhancing Spatial Perception and Reasoning in Vision-Language Models via Geometric Surrogate Tasks](https://mlsanity.com/paper/2509.24473/) (2025)
- [SI-Score: An image dataset for fine-grained analysis of robustness to object location, rotation and size](https://mlsanity.com/paper/2104.04191/) (2021)
- [Why Representation Engineering Works: A Theoretical and Empirical Study in Vision-Language Models](https://mlsanity.com/paper/2503.22720/) (2025)
- [RotBench: Evaluating Multimodal Large Language Models on Identifying Image Rotation](https://mlsanity.com/paper/2508.13968/) (2025)

## Abstract (arXiv)

Multimodal Large Language Models (MLLMs) struggle with tasks that require reasoning about 2D object orientation in images, as documented in prior work. Tong et al. and Nichols et al. hypothesize that these failures originate in the visual encoder, since commonly used encoders such as CLIP and SigLIP are trained for image-text semantic alignment rather than geometric reasoning. We design a controlled empirical protocol to test this claim by measuring whether rotations can be recovered from encoder representations. In particular, we examine SigLIP and ViT features from LLaVA OneVision and Qwen2.5-VL-7B-Instruct models, respectively, using full images, and examine CLIP representations in LLaVA 1.5 and 1.6 using rotated foreground patches against natural background images. Our null hypothesis is that orientation information is not preserved in the encoder embeddings and we test this by training linear regressors to predict object orientation from encoded features. Contrary to the hypothesis, we find that orientation information is recoverable from encoder representations: simple linear models accurately predict object orientations from embeddings. This contradicts the assumption that MLLM orientation failures originate in the visual encoder. Having rejected the accepted hypothesis that MLLMs struggle with 2D orientation tasks because of visual encoder limitations, we still don't know why they fail. Although a full explanation is beyond the scope of this paper, we show that although present, orientation information is spread diffusely across tens of thousands of features. This may or may not be while MLLMs fail to exploit the available orientation information.