Research map / Video / Video-language understanding
Video SimpleQA: Towards Factuality Evaluation in Large Video Language Models
Meng Cao, Pengfei Hu, Yingyao Wang, Jihao Gu, Haoran Tang, Haoze Zhao, Chen Wang, Jiahua Dong, Wangbo Yu, Ge Zhang, Xiang Li, Ian Reid and 1 more
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 retrieval-augmented generation (LlamaIndex with web search) (ablation) | Claude Sonnet 4 · Video SimpleQA | F-score: 35.6 -> 58.8 | Improved |
| add retrieval-augmented generation (ablation) | GPT-4o · Video SimpleQA | F-score: 49.3 -> 61.3 | Improved |
| add retrieval-augmented generation (ablation) | Qwen-VL-Max · Video SimpleQA | F-score: 39.9 -> 57.0 | Improved |
| add retrieval-augmented generation (ablation) | o3 · Video SimpleQA | F-score: 66.3 -> 69.0 | Improved |
| add retrieval-augmented generation (ablation) | Gemini 2.5 Pro · Video SimpleQA | F-score: 62.6 -> 66.2 | Improved |
| add retrieval-augmented generation (ablation) | o3 · Video SimpleQA | inference time (min): 27.8 -> 54.7 | Hurt |
| add retrieval-augmented generation (ablation) | Qwen-VL-Max · Video SimpleQA | inference time (min): 24.2 -> 61.2 | Hurt |
| ask for final multi-hop answer instead of per-hop sub-questions | GPT-4o · Video SimpleQA 4-hop | F-score: 85.4 (QA2) -> 47.0 (multi-hop) | Hurt |
| first hop requires video object/event recognition versus later text-only hops | Claude Sonnet 4 · Video SimpleQA 4-hop | F-score: 43.5 (QA1) vs 85.7 (QA2) | Hurt |
| scale model size within Qwen2.5-VL (7B to 72B) | Qwen2.5-VL · Video SimpleQA | F-score: 25.3 -> 39.5 | Improved |
| human access to external resources (open-book vs closed-book) | Human annotators · Video SimpleQA (200 sampled) | F-score: 66.5 -> 90.6 | Improved |
Closest papers by meaning
- MMBench-Video: A Long-Form Multi-Shot Benchmark for Holistic Video Understanding (NeurIPS 2024 · similarity 0.91)
- LongVideoBench: A Benchmark for Long-context Interleaved Video-Language Understanding (NeurIPS 2024 · similarity 0.90)
- RTV-Bench: Benchmarking MLLM Continuous Perception, Understanding and Reasoning through Real-Time Video (NeurIPS 2025 · similarity 0.89)
- Q-Bench-Video: Benchmark the Video Quality Understanding of LMMs (CVPR 2025 · similarity 0.89)
- SAMA: Towards Multi-Turn Referential Grounded Video Chat with Large Language Models (NeurIPS 2025 · similarity 0.89)
- Needle In A Video Haystack: A Scalable Synthetic Evaluator for Video MLLMs (ICLR 2025 · similarity 0.89)
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.
