AI Paper Atlas

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

CVPR 2026 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
add zero-shot chain-of-thought prompting (ablation)Qwen2.5-VL-72B · Flat-Pack Benchmicro avg accuracy: 40.19 -> 39.20Hurt
add self-consistency CoT with 5 samples (ablation)Qwen2.5-VL-72B · Flat-Pack Benchmicro avg accuracy: 40.19 -> 32.23Hurt
remove the video and give image-only prompt (ablation)Qwen2.5-VL-72B · Flat-Pack Benchmicro avg accuracy: 40.20 -> 31.40Hurt
remove the video and give image-only prompt (ablation)Qwen2.5-VL-72B · Flat-Pack Bench Trackaccuracy: 45.53 -> 21.01Hurt
remove the video and give image-only prompt (ablation)Qwen2.5-VL-72B · Flat-Pack Bench Mateaccuracy: 35.63 -> 41.38Improved
remove the video and give image-only prompt (ablation)Qwen2.5-VL-72B · Flat-Pack Bench TLocaccuracy: 30.10 -> 34.95Improved
shuffle parts in image-only prompt (ablation)Qwen2.5-VL-72B · Flat-Pack Benchmicro avg accuracy: 40.20 -> 29.96Hurt
use trimmed versus key-frame video or different visual prompt types (ablation)Qwen2.5-VL-72B · Flat-Pack BenchaccuracyNo clear effect
use greedy high-contrast mask colors instead of random colors (ablation)Qwen2.5-VL-72B · Flat-Pack BenchaccuracyNo clear effect
render label, boundary and mask together instead of subsets (ablation)Qwen2.5-VL-72B · Flat-Pack BenchaccuracyImproved
select question-relevant frames with GenSGemini 2.5 Pro · Flat-Pack Benchmicro avg accuracy: 33.72 -> 25.58Hurt
use ArrowRL temporal training instead of base checkpointQwen2.5-VL-7B · Flat-Pack Benchmicro avg accuracy: 30.23 -> 30.56Mixed

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy