Research map / Video / Video-language understanding
Flat-Pack Bench: Evaluating Spatio-Temporal Understanding in Large Vision-Language Models through Furniture Assembly
Aditya Chetan, Eric Cai, Peeyush Kushwaha, Bharath Raj Nagoor Kani, Utkarsh Mall, Qianqian Wang, Noah Snavely, Bharath Hariharan
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add zero-shot chain-of-thought prompting (ablation) | Qwen2.5-VL-72B · Flat-Pack Bench | micro avg accuracy: 40.19 -> 39.20 | Hurt |
| add self-consistency CoT with 5 samples (ablation) | Qwen2.5-VL-72B · Flat-Pack Bench | micro avg accuracy: 40.19 -> 32.23 | Hurt |
| remove the video and give image-only prompt (ablation) | Qwen2.5-VL-72B · Flat-Pack Bench | micro avg accuracy: 40.20 -> 31.40 | Hurt |
| remove the video and give image-only prompt (ablation) | Qwen2.5-VL-72B · Flat-Pack Bench Track | accuracy: 45.53 -> 21.01 | Hurt |
| remove the video and give image-only prompt (ablation) | Qwen2.5-VL-72B · Flat-Pack Bench Mate | accuracy: 35.63 -> 41.38 | Improved |
| remove the video and give image-only prompt (ablation) | Qwen2.5-VL-72B · Flat-Pack Bench TLoc | accuracy: 30.10 -> 34.95 | Improved |
| shuffle parts in image-only prompt (ablation) | Qwen2.5-VL-72B · Flat-Pack Bench | micro avg accuracy: 40.20 -> 29.96 | Hurt |
| use trimmed versus key-frame video or different visual prompt types (ablation) | Qwen2.5-VL-72B · Flat-Pack Bench | accuracy | No clear effect |
| use greedy high-contrast mask colors instead of random colors (ablation) | Qwen2.5-VL-72B · Flat-Pack Bench | accuracy | No clear effect |
| render label, boundary and mask together instead of subsets (ablation) | Qwen2.5-VL-72B · Flat-Pack Bench | accuracy | Improved |
| select question-relevant frames with GenS | Gemini 2.5 Pro · Flat-Pack Bench | micro avg accuracy: 33.72 -> 25.58 | Hurt |
| use ArrowRL temporal training instead of base checkpoint | Qwen2.5-VL-7B · Flat-Pack Bench | micro avg accuracy: 30.23 -> 30.56 | Mixed |
Closest papers by meaning
- OST-Bench: Evaluating the Capabilities of MLLMs in Online Spatio-temporal Scene Understanding (NeurIPS 2025 · similarity 0.89)
- PerceptionLM: Open-Access Data and Models for Detailed Visual Understanding (NeurIPS 2025 · similarity 0.89)
- MMBench-Video: A Long-Form Multi-Shot Benchmark for Holistic Video Understanding (NeurIPS 2024 · similarity 0.88)
- InternSpatial: A Comprehensive Dataset for Spatial Reasoning in Vision-Language Models (ICLR 2026 · similarity 0.88)
- VideoRefer Suite: Advancing Spatial-Temporal Object Understanding with Video LLM (CVPR 2025 · similarity 0.88)
- Spatial Understanding from Videos: Structured Prompts Meet Simulation Data (NeurIPS 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
