# Benchmarking and Mechanistic Analysis of Vision-Language Models for Cross-Depiction Assembly Instruction Alignment

arXiv: https://arxiv.org/abs/2604.00913 · HTML: https://mlsanity.com/paper/2604.00913/

## Key facts

- No citing papers in the map yet (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 16 of its 31 references that are in the map.
- First posted on arXiv 1 Apr 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, 16 of 16)

- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [Similarity of Neural Network Representations Revisited](https://mlsanity.com/paper/1905.00414/) (2019)
- [Understanding intermediate layers using linear classifier probes](https://mlsanity.com/paper/1610.01644/) (2016)
- [Assembly101: A Large-Scale Multi-View Video Dataset for Understanding Procedural Activities](https://mlsanity.com/paper/2203.14712/) (2022)
- [HoloAssist: an Egocentric Human Interaction Dataset for Interactive AI Assistants in the Real World](https://mlsanity.com/paper/2309.17024/) (2023)
- [The IKEA ASM Dataset: Understanding People Assembling Furniture through Actions, Objects and Pose](https://mlsanity.com/paper/2007.00394/) (2020)
- [The Hidden Life of Tokens: Reducing Hallucination of Large Vision-Language Models via Visual Information Steering](https://mlsanity.com/paper/2502.03628/) (2025)
- [CaptainCook4D: A Dataset for Understanding Errors in Procedural Activities](https://mlsanity.com/paper/2312.14556/) (2023)
- [What Do VLMs NOTICE? A Mechanistic Interpretability Pipeline for Gaussian-Noise-free Text-Image Corruption and Evaluation](https://mlsanity.com/paper/2406.16320/) (2024)
- [IKEA-Manual: Seeing Shape Assembly Step by Step](https://mlsanity.com/paper/2302.01881/) (2023)
- [Manual2Skill: Learning to Read Manuals and Acquire Robotic Skills for Furniture Assembly Using Vision-Language Models](https://mlsanity.com/paper/2502.10090/) (2025)
- [IKEA Manuals at Work: 4D Grounding of Assembly Instructions on Internet Videos](https://mlsanity.com/paper/2411.11409/) (2024)
- [The Cross-Depiction Problem: Computer Vision Algorithms for Recognising Objects in Artwork and in Photographs](https://mlsanity.com/paper/1505.00110/) (2015)
- [More Images, More Problems? A Controlled Analysis of VLM Failure Modes](https://mlsanity.com/paper/2601.07812/) (2026)
- [LEGO Co-builder: Exploring Fine-Grained Vision-Language Modeling for Multimodal LEGO Assembly Assistants](https://mlsanity.com/paper/2507.05515/) (2025)
- [TechING: Towards Real World Technical Image Understanding via VLMs](https://mlsanity.com/paper/2601.18238/) (2026)

## Abstract (arXiv)

2D assembly diagrams are often abstract and hard to follow, creating a need for intelligent assistants that can monitor progress, detect errors, and provide step-by-step guidance. In mixed reality settings, such systems must recognize completed and ongoing steps from the camera feed and align them with the diagram instructions. Vision Language Models (VLMs) show promise for this task, but face a depiction gap because assembly diagrams and video frames share few visual features. To systematically assess this gap, we construct IKEA-Bench, a benchmark of 1,623 questions across 6 task types on 29 IKEA furniture products, and evaluate 19 VLMs (2B-38B) under three alignment strategies. Our key findings: (1) assembly instruction understanding is recoverable via text, but text simultaneously degrades diagram-to-video alignment; (2) architecture family predicts alignment accuracy more strongly than parameter count; (3) video understanding remains a hard bottleneck unaffected by strategy. A three-level mechanistic analysis further reveals that diagrams and video occupy disjoint ViT subspaces, and that adding text shifts models from visual to text-driven reasoning. These results identify visual encoding as the primary target for improving cross-depiction robustness. Project page: https://ryenhails.github.io/IKEA-Bench/