---
title: "CoV: Chain-of-View Prompting for Spatial Reasoning (2026)"
description: "The authors propose Chain-of-View (CoV) prompting, a training-free procedure applied at inference time that lets a VLM… 49 references."
authors: ["Haoyu Zhao", "Akide Liu", "Zeyu Zhang", "Weijie Wang", "Feng Chen", "Ruihan Zhu", "Gholamreza Haffari", "Bohan Zhuang"]
year: 2026
venue: "Findings of the Association for Computational Linguistics: ACL 2026"
source: "https://aclanthology.org/2026.findings-acl.1623/"
doi: "10.18653/v1/2026.findings-acl.1623"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/zhao-2026-cov-chain-view-prompting-spatial/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/zhao-2026-cov-chain-view-prompting-spatial/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/7/) / CoV: Chain-of-View Prompting for Spatial Reasoning

# CoV: Chain-of-View Prompting for Spatial Reasoning

**Authors:** Haoyu Zhao, Akide Liu, [Zeyu Zhang](https://md.research.elamur.ai/free/en/authors/zeyu-zhang/), [Weijie Wang](https://md.research.elamur.ai/free/en/authors/weijie-wang/), [Feng Chen](https://md.research.elamur.ai/free/en/authors/feng-chen/), Ruihan Zhu, [Gholamreza Haffari](https://md.research.elamur.ai/free/en/authors/gholamreza-haffari/) and Bohan Zhuang · **Year:** 2026 · **Venue:** [Findings of the Association for Computational Linguistics: ACL 2026](https://md.research.elamur.ai/free/en/venues/findings/) ([Findings of the Association for Computational Linguistics 2026](https://md.research.elamur.ai/free/en/venues/findings/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.findings-acl.1623/) · [PDF](https://aclanthology.org/2026.findings-acl.1623.pdf) · [DOI 10.18653/v1/2026.findings-acl.1623](https://doi.org/10.18653/v1/2026.findings-acl.1623) · [OpenAlex](https://openalex.org/W7166825444)

**Citations:** 2 citations (OpenAlex, 2026-10-14)

## Summary
Vision–language models (VLMs) used for embodied question answering (EQA) in 3D scenes usually receive a small, fixed set of frames. Evidence that is hidden or spread across a room may therefore never reach the model. The authors propose Chain-of-View (CoV) prompting, a training-free procedure applied at inference time that lets a VLM decide where to look next.

CoV works in two stages:
- **Coarse stage:** a View Selection agent, which also sees a bird's-eye image of the scene, keeps only a few anchor frames that match the question.
- **Fine stage:** a CoV agent alternates between reasoning and discrete camera commands, such as rotating or moving forward. Each command becomes a rigid camera transform, and a new view is rendered from the 3D scene and passed back to the model. This repeats until the model judges it has enough evidence or reaches a step limit.

On OpenEQA, CoV clearly beats the unmodified VLMs and improves further when more action steps are allowed. It also performs well on ScanQA and SQA3D. The work shows that spatial reasoning can be improved in a model-agnostic way without any fine-tuning.

## Key points
- **OpenEQA (LLM-Match):** CoV was tested with Qwen3-VL-Flash, GLM-4.6V, GPT-4o-mini and Gemini-2.5-Flash. The average gain is +11.98% (relative), and the largest is +13.62% on Qwen3-VL-Flash (52.65 → 59.82).
- **Test-time scaling:** raising the minimum number of action steps (up to 7) adds +2.54% on average, peaking at +3.73% on Gemini-2.5-Flash.
- **Other benchmarks:** 116 CIDEr and 31.9 EM@1 on ScanQA; 51.1 EM@1 on SQA3D.
- **Ablations (Qwen3-VL-Flash, 3 steps, CoV = 59.03):**
  - Dropping coarse view selection lowers scores by 2.71% on average across the four models.
  - K = 5 anchor views works best (K = 3: 48.16; K = 7: 52.55).
  - View selection with a static VLM scores only 35.77.
  - Giving a static VLM the newly explored views scores 47.69.
- **Limitations:** there is no explicit stopping rule (termination relies on the VLM's own judgment), results depend on the quality of view selection, and the method may struggle in dynamic or cluttered scenes. Code is at github.com/ziplab/CoV.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

📚 Shelves (5): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 16,039 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 15,891 works · [👁️ Computer Vision and Pattern Recognition](https://md.research.elamur.ai/free/en/shelves/computer-science/computer-vision-and-pattern-recognition/) · 2,869 works

🧭 Topics (2): [🧠 Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/) · 3,320 works · [🧮 Reasoning](https://md.research.elamur.ai/free/en/topics/reasoning/) · 1,426 works

## Cite

Cite as: Haoyu Zhao, Akide Liu, Zeyu Zhang, Weijie Wang, Feng Chen, Ruihan Zhu, Gholamreza Haffari and Bohan Zhuang. 2026. CoV: Chain-of-View Prompting for Spatial Reasoning. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{zhao2026cov,
  title = {{CoV: Chain-of-View Prompting for Spatial Reasoning}},
  author = {Haoyu Zhao, Akide Liu, Zeyu Zhang, Weijie Wang, Feng Chen, Ruihan Zhu, Gholamreza Haffari and Bohan Zhuang},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.1623},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.1623}
}
```

- [← previous Confidence over Time: Confidence Calibration with Temporal Logic for Large Language Model Reasoning](https://md.research.elamur.ai/free/en/papers/mao-2026-confidence-over-time-confidence-calibration/)
- [next → Cross-layer Attention Sharing for Pre-trained Large Language Models](https://md.research.elamur.ai/free/en/papers/mu-2026-cross-layer-attention-sharing-pre/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/7/)

## References

49 references, 15 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Panos Achlioptas, Ahmed Abdelreheem, Fei Xia, Mohamed Elhoseiny, and Leonidas Guibas. 2020. Referit3d: Neural listeners for fine-grained 3d object identification in real-world scenes. 16th European Conference on Computer Vision (ECCV). [find](https://scholar.google.com/scholar?q=Panos+Achlioptas%2C+Ahmed+Abdelreheem%2C+Fei+Xia%2C+Mohamed+Elhoseiny%2C+and+Leonidas+Guibas%2E+2020%2E+Referit3d%3A+Neural+listeners+for+fine%2Dgrained+3d+object+identification+in+real%2Dworld+scenes%2E+16th+European+Co)
2. Daichi Azuma, Taiki Miyanishi, Shuhei Kurita, and Motoaki Kawanabe. 2022. Scanqa: 3d question answering for spatial scene understanding. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 19107– 19117. [find](https://scholar.google.com/scholar?q=Daichi+Azuma%2C+Taiki+Miyanishi%2C+Shuhei+Kurita%2C+and+Motoaki+Kawanabe%2E+2022%2E+Scanqa%3A+3d+question+answering+for+spatial+scene+understanding%2E+In+Proceedings+of+the+IEEE%2FCVF+Conference+on+Computer+Vision+an)
3. Shuai Bai, Yuxuan Cai, Ruizhe Chen, Keqin Chen, Xionghui Chen, Zesen Cheng, Lianghao Deng, Wei Ding, Chang Gao, Chunjiang Ge, Wenbin Ge, Zhifang Guo, Qidong Huang, Jie Huang, Fei Huang, Binyuan Hui, Shutong Jiang, Zhaohai Li, Mingsheng Li, and 45 others. 2025. [Qwen3-vl technical report](https://md.research.elamur.ai/free/en/papers/bai-2025-qwen3-vl-technical-report/). Preprint, ar…
4. Tom B Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, and 1 others. 2020. [Language models are few-shot learners](https://md.research.elamur.ai/free/en/papers/brown-2020-language-models-few-shot-learners/). Advances in Neural Information Processing Systems, 33:1877–1901.
5. Dave Zhenyu Chen, Angel X Chang, and Matthias Nießner. 2020. Scanrefer: 3d object localization in rgb-d scans using natural language. 16th European Conference on Computer Vision (ECCV). [find](https://scholar.google.com/scholar?q=Dave+Zhenyu+Chen%2C+Angel+X+Chang%2C+and+Matthias+Nie%C3%9Fner%2E+2020%2E+Scanrefer%3A+3d+object+localization+in+rgb%2Dd+scans+using+natural+language%2E+16th+European+Conference+on+Computer+Vision+%28ECCV%29%2E)
6. Dave Zhenyu Chen, Ali Gholami, Matthias Nießner, and Angel X. Chang. 2022. D3net: A unified speaker-listener architecture for 3d dense captioning and visual grounding. In Proceedings of the European Conference on Computer Vision (ECCV). [find](https://scholar.google.com/scholar?q=Dave+Zhenyu+Chen%2C+Ali+Gholami%2C+Matthias+Nie%C3%9Fner%2C+and+Angel+X%2E+Chang%2E+2022%2E+D3net%3A+A+unified+speaker%2Dlistener+architecture+for+3d+dense+captioning+and+visual+grounding%2E+In+Proceedings+of+the+European+C)
7. Sijin Chen, Xin Chen, Chi Zhang, Mingsheng Li, Gang Yu, Hao Fei, Hongyuan Zhu, Jiayuan Fan, and Tao Chen. 2024. Ll3da: Visual interactive instruction tuning for omni-3d understanding reasoning and planning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR),… [find](https://scholar.google.com/scholar?q=Sijin+Chen%2C+Xin+Chen%2C+Chi+Zhang%2C+Mingsheng+Li%2C+Gang+Yu%2C+Hao+Fei%2C+Hongyuan+Zhu%2C+Jiayuan+Fan%2C+and+Tao+Chen%2E+2024%2E+Ll3da%3A+Visual+interactive+instruction+tuning+for+omni%2D3d+understanding+reasoning+and+pla)
8. Sijin Chen, Hongyuan Zhu, Xin Chen, Yinjie Lei, Gang Yu, and Tao Chen. 2023a. End-to-end 3d dense captioning with vote2cap-detr. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 11124–11133. [find](https://scholar.google.com/scholar?q=Sijin+Chen%2C+Hongyuan+Zhu%2C+Xin+Chen%2C+Yinjie+Lei%2C+Gang+Yu%2C+and+Tao+Chen%2E+2023a%2E+End%2Dto%2Dend+3d+dense+captioning+with+vote2cap%2Ddetr%2E+In+Proceedings+of+the+IEEE%2FCVF+Conference+on+Computer+Vision+and+Patter)
9. Zhe Chen, Weiyun Wang, Yue Cao, Yangzhou Liu, Zhangwei Gao, Erfei Cui, Jinguo Zhu, Shenglong Ye, Hao Tian, Zhaoyang Liu, and 1 others. 2023b. [Internvl: Scaling up vision foundation models and aligning for generic vision-language tasks](https://md.research.elamur.ai/free/en/papers/chen-2023-internvl-scaling-up-vision/). arXiv preprint arXiv:2312.14238.
10. Zhenyu Chen, Ali Gholami, Matthias Nießner, and Angel X. Chang. 2021. Scan2cap: Context-aware dense captioning in rgb-d scans. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 3195–3204. [find](https://scholar.google.com/scholar?q=Zhenyu+Chen%2C+Ali+Gholami%2C+Matthias+Nie%C3%9Fner%2C+and+Angel+X%2E+Chang%2E+2021%2E+Scan2cap%3A+Context%2Daware+dense+captioning+in+rgb%2Dd+scans%2E+In+Proceedings+of+the+IEEE%2FCVF+Conference+on+Computer+Vision+and+Pattern+)
11. Gheorghe Comanici, Eric Bieber, Mike Schaekermann, Ice Pasupat, Noveen Sachdeva, Inderjit Dhillon, Marcel Blistein, Ori Ram, Dan Zhang, Evan Rosen, Luke Marris, Sam Petulla, Colin Gaffney, Asaf Aharoni, Nathan Lintz, Tiago Cardal Pais, Henrik Jacobsson, Idan Szpektor, Nan-Jiang Jiang, and 3416 [other…](https://md.research.elamur.ai/free/en/papers/comanici-2025-gemini-2-5-pushing-frontier/)
12. Angela Dai, Angel X Chang, Manolis Savva, Maciej Halber, Thomas Funkhouser, and Matthias Nießner. 2017. Scannet: Richly-annotated 3d reconstructions of indoor scenes. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 5828–5839. [find](https://scholar.google.com/scholar?q=Angela+Dai%2C+Angel+X+Chang%2C+Manolis+Savva%2C+Maciej+Halber%2C+Thomas+Funkhouser%2C+and+Matthias+Nie%C3%9Fner%2E+2017%2E+Scannet%3A+Richly%2Dannotated+3d+reconstructions+of+indoor+scenes%2E+In+Proceedings+of+the+IEEE+confer)
13. Rao Fu, Jingyu Liu, Xilun Chen, Yixin Nie, and Wenhan Xiong. 2024. Scene-llm: Extending language model for 3d visual understanding and reasoning. arXiv preprint arXiv:2403.11401. — [arXiv:2403.11401](https://arxiv.org/abs/2403.11401) [find](https://scholar.google.com/scholar?q=Rao+Fu%2C+Jingyu+Liu%2C+Xilun+Chen%2C+Yixin+Nie%2C+and+Wenhan+Xiong%2E+2024%2E+Scene%2Dllm%3A+Extending+language+model+for+3d+visual+understanding+and+reasoning%2E+arXiv+preprint+arXiv%3A2403%2E11401%2E+%E2%80%94+%5BarXiv%3A2403%2E11401%5D%28)
14. Petko Georgiev and 1 others. 2024. [Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context](https://md.research.elamur.ai/free/en/papers/gemini-team-2024-gemini-1-5-unlocking-multimodal/). arXiv preprint arXiv:2403.05530.
15. Yining Hong, Haoyu Zhen, Peihao Chen, Shuhong Zheng, Yilun Du, Zhenfang Chen, and Chuang Gan. 2023. 3d-llm: Injecting the 3d world into large language models. In Advances in Neural Information Processing Systems, volume 36, pages 20482–20494. Curran Associates, Inc. [find](https://scholar.google.com/scholar?q=Yining+Hong%2C+Haoyu+Zhen%2C+Peihao+Chen%2C+Shuhong+Zheng%2C+Yilun+Du%2C+Zhenfang+Chen%2C+and+Chuang+Gan%2E+2023%2E+3d%2Dllm%3A+Injecting+the+3d+world+into+large+language+models%2E+In+Advances+in+Neural+Information+Process)
16. Haifeng Huang, Yilun Chen, Zehan Wang, Rongjie Huang, Runsen Xu, Tai Wang, Luping Liu, Xize Cheng, Yang Zhao, Jiangmiao Pang, and Zhou Zhao. 2024a. Chat-scene: Bridging 3d scene and large language models with object identifiers. Preprint, arXiv:2312.08168. — [arXiv:2312.08168](https://arxiv.org/abs/2312.08168) [find](https://scholar.google.com/scholar?q=Haifeng+Huang%2C+Yilun+Chen%2C+Zehan+Wang%2C+Rongjie+Huang%2C+Runsen+Xu%2C+Tai+Wang%2C+Luping+Liu%2C+Xize+Cheng%2C+Yang+Zhao%2C+Jiangmiao+Pang%2C+and+Zhou+Zhao%2E+2024a%2E+Chat%2Dscene%3A+Bridging+3d+scene+and+large+language+mod)
17. Haifeng Huang, Yilun Chen, Zehan Wang, Rongjie Huang, Runsen Xu, Tai Wang, Luping Liu, Xize Cheng, Yang Zhao, Jiangmiao Pang, and Zhou Zhao. 2024b. Chat-scene: Bridging 3d scene and large language models with object identifiers. Preprint, arXiv:2312.08168. — [arXiv:2312.08168](https://arxiv.org/abs/2312.08168) [find](https://scholar.google.com/scholar?q=Haifeng+Huang%2C+Yilun+Chen%2C+Zehan+Wang%2C+Rongjie+Huang%2C+Runsen+Xu%2C+Tai+Wang%2C+Luping+Liu%2C+Xize+Cheng%2C+Yang+Zhao%2C+Jiangmiao+Pang%2C+and+Zhou+Zhao%2E+2024b%2E+Chat%2Dscene%3A+Bridging+3d+scene+and+large+language+mod)
18. Haifeng Huang, Zehan Wang, Rongjie Huang, Luping Liu, Xize Cheng, Yang Zhao, Tao Jin, and Zhou Zhao. 2023. Chat-3d v2: Bridging 3d scene and large language models with object identifiers. arXiv preprint arXiv:2312.08168. — [arXiv:2312.08168](https://arxiv.org/abs/2312.08168) [find](https://scholar.google.com/scholar?q=Haifeng+Huang%2C+Zehan+Wang%2C+Rongjie+Huang%2C+Luping+Liu%2C+Xize+Cheng%2C+Yang+Zhao%2C+Tao+Jin%2C+and+Zhou+Zhao%2E+2023%2E+Chat%2D3d+v2%3A+Bridging+3d+scene+and+large+language+models+with+object+identifiers%2E+arXiv+prepri)
19. Jiangyong Huang, Silong Yong, Xiaojian Ma, Xiongkun Linghu, Puhao Li, Yan Wang, Qing Li, Song-Chun Zhu, Baoxiong Jia, and Siyuan Huang. 2024c. An embodied generalist agent in 3d world. Preprint, arXiv:2311.12871. — [arXiv:2311.12871](https://arxiv.org/abs/2311.12871) [find](https://scholar.google.com/scholar?q=Jiangyong+Huang%2C+Silong+Yong%2C+Xiaojian+Ma%2C+Xiongkun+Linghu%2C+Puhao+Li%2C+Yan+Wang%2C+Qing+Li%2C+Song%2DChun+Zhu%2C+Baoxiong+Jia%2C+and+Siyuan+Huang%2E+2024c%2E+An+embodied+generalist+agent+in+3d+world%2E+Preprint%2C+arXiv)
20. Ting Huang, Zeyu Zhang, Yemin Wang, and Hao Tang. 2025a. 3d coca: Contrastive learners are 3d captioners. arXiv preprint arXiv:2504.09518. — [arXiv:2504.09518](https://arxiv.org/abs/2504.09518) [find](https://scholar.google.com/scholar?q=Ting+Huang%2C+Zeyu+Zhang%2C+Yemin+Wang%2C+and+Hao+Tang%2E+2025a%2E+3d+coca%3A+Contrastive+learners+are+3d+captioners%2E+arXiv+preprint+arXiv%3A2504%2E09518%2E+%E2%80%94+%5BarXiv%3A2504%2E09518%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F2504%2E09518%29)
21. Ting Huang, Zeyu Zhang, Ruicheng Zhang, and Yang Zhao. 2025b. Dc-scene: Data-centric learning for 3d scene understanding. arXiv preprint arXiv:2505.15232. — [arXiv:2505.15232](https://arxiv.org/abs/2505.15232) [find](https://scholar.google.com/scholar?q=Ting+Huang%2C+Zeyu+Zhang%2C+Ruicheng+Zhang%2C+and+Yang+Zhao%2E+2025b%2E+Dc%2Dscene%3A+Data%2Dcentric+learning+for+3d+scene+understanding%2E+arXiv+preprint+arXiv%3A2505%2E15232%2E+%E2%80%94+%5BarXiv%3A2505%2E15232%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F25)
22. Jie Lei, Linjie Li, Luowei Zhou, Zhe Gan, Tamara L. Berg, Mohit Bansal, and Jingjing Liu. 2021. Less is more: Clipbert for video-and-language learning via sparse sampling. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 7331–7341. [find](https://scholar.google.com/scholar?q=Jie+Lei%2C+Linjie+Li%2C+Luowei+Zhou%2C+Zhe+Gan%2C+Tamara+L%2E+Berg%2C+Mohit+Bansal%2C+and+Jingjing+Liu%2E+2021%2E+Less+is+more%3A+Clipbert+for+video%2Dand%2Dlanguage+learning+via+sparse+sampling%2E+In+Proceedings+of+the+IEEE%2FC)
23. Bo Li, Yuanhan Zhang, Dong Guo, Renrui Zhang, Feng Li, Hao Zhang, Kaichen Zhang, Peiyuan Zhang, Yanwei Li, Ziwei Liu, and Chunyuan Li. 2024a. [Llava-onevision: Easy visual task transfer](https://md.research.elamur.ai/free/en/papers/li-2024-llava-onevision-easy-visual-task/). Preprint, arXiv:2408.03326.
24. Kunchang Li, Yali Wang, Yinan He, Yizhuo Li, Yi Wang, Yi Liu, Zun Wang, Jilan Xu, Guo Chen, Ping Luo, Limin Wang, and Yu Qiao. 2024b. [Mvbench: A comprehensive multi-modal video understanding benchmark](https://md.research.elamur.ai/free/en/papers/li-2023-mvbench-comprehensive-multi-modal-video/). In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages…
25. Yifan Li, Yikang Wang, Yang Zhao, Ziang Zhang, and Zhou Zhao. 2024c. Video-3d llm: Learning position-aware video representation for 3d scene understanding. arXiv preprint arXiv:2412.00493. — [arXiv:2412.00493](https://arxiv.org/abs/2412.00493) [find](https://scholar.google.com/scholar?q=Yifan+Li%2C+Yikang+Wang%2C+Yang+Zhao%2C+Ziang+Zhang%2C+and+Zhou+Zhao%2E+2024c%2E+Video%2D3d+llm%3A+Learning+position%2Daware+video+representation+for+3d+scene+understanding%2E+arXiv+preprint+arXiv%3A2412%2E00493%2E+%E2%80%94+%5BarXiv%3A24)
26. Xiaojian Ma, Silong Yong, Zilong Zheng, Qing Li, Yitao Liang, Song-Chun Zhu, and Siyuan Huang. 2023. Sqa3d: Situated question answering in 3d scenes. In Proceedings of the International Conference on Learning Representations (ICLR). [find](https://scholar.google.com/scholar?q=Xiaojian+Ma%2C+Silong+Yong%2C+Zilong+Zheng%2C+Qing+Li%2C+Yitao+Liang%2C+Song%2DChun+Zhu%2C+and+Siyuan+Huang%2E+2023%2E+Sqa3d%3A+Situated+question+answering+in+3d+scenes%2E+In+Proceedings+of+the+International+Conference+on+)
27. Arjun Majumdar, Anurag Ajay, Xiaohan Zhang, Pranav Putta, Sriram Yenamandra, Mikael Henaff, Sneha Silwal, Paul Mcvay, Oleksandr Maksymets, Sergio Arnaud, Karmesh Yadav, Qiyang Li, Ben Newman, Mohit Sharma, Vincent Berges, Shiqi Zhang, Pulkit Agrawal, Yonatan Bisk, Dhruv Batra, and 5 others. 2024. Op… [find](https://scholar.google.com/scholar?q=Arjun+Majumdar%2C+Anurag+Ajay%2C+Xiaohan+Zhang%2C+Pranav+Putta%2C+Sriram+Yenamandra%2C+Mikael+Henaff%2C+Sneha+Silwal%2C+Paul+Mcvay%2C+Oleksandr+Maksymets%2C+Sergio+Arnaud%2C+Karmesh+Yadav%2C+Qiyang+Li%2C+Ben+Newman%2C+Mohit+Sh)
28. Stephen McKenna and Jacob Carse. 2024. Calibrating where it matters: Constrained temperature scaling. arXiv preprint arXiv:2406.11456. — [arXiv:2406.11456](https://arxiv.org/abs/2406.11456) [find](https://scholar.google.com/scholar?q=Stephen+McKenna+and+Jacob+Carse%2E+2024%2E+Calibrating+where+it+matters%3A+Constrained+temperature+scaling%2E+arXiv+preprint+arXiv%3A2406%2E11456%2E+%E2%80%94+%5BarXiv%3A2406%2E11456%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F2406%2E11456%29)
29. V Team, Wenyi Hong, Wenmeng Yu, Xiaotao Gu, Guo Wang, Guobing Gan, Haomiao Tang, Jiale Cheng, Ji Qi, Junhui Ji, Lihang Pan, Shuaiqi Duan, Weihan Wang, Yan Wang, Yean Cheng, Zehai He, Zhe Su, Zhen Yang, Ziyang Pan, and 69 others. 2025. [Glm-4.5v and glm-4.1v-thinking: Towards versatile multimodal reas…](https://md.research.elamur.ai/free/en/papers/team-2025-v-glm-4-5v-glm/)
30. Wentao Mo and Yang Liu. 2024. Bridging the gap between 2d and 3d visual question answering: A fusion approach for 3d vqa. arXiv preprint arXiv:2402.15933. — [arXiv:2402.15933](https://arxiv.org/abs/2402.15933) [find](https://scholar.google.com/scholar?q=Wentao+Mo+and+Yang+Liu%2E+2024%2E+Bridging+the+gap+between+2d+and+3d+visual+question+answering%3A+A+fusion+approach+for+3d+vqa%2E+arXiv+preprint+arXiv%3A2402%2E15933%2E+%E2%80%94+%5BarXiv%3A2402%2E15933%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F24)
31. Heng Wang, Jianhui Yu, and Weidong Cai. 2022. Spacap3d: Spatiality-guided transformer for 3d dense captioning on point clouds. In Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI). [find](https://scholar.google.com/scholar?q=Heng+Wang%2C+Jianhui+Yu%2C+and+Weidong+Cai%2E+2022%2E+Spacap3d%3A+Spatiality%2Dguided+transformer+for+3d+dense+captioning+on+point+clouds%2E+In+Proceedings+of+the+International+Joint+Conference+on+Artificial+Intell)
32. Niklas Muennighoff, Zitong Yang, Weijia Shi, Xiang Lisa Li, Li Fei-Fei, Hannaneh Hajishirzi, Luke Zettlemoyer, Percy Liang, Emmanuel Candès, and Tatsunori Hashimoto. 2025a. [s1: Simple test-time scaling](https://md.research.elamur.ai/free/en/papers/muennighoff-2025-s1-simple-test-time-scaling/). arXiv preprint arXiv:2501.19393.
33. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Brian Ichter, Fei Xia, and 1 others. 2022. [Chain-of-thought prompting elicits reasoning in large language models](https://md.research.elamur.ai/free/en/papers/wei-2022-chain-thought-prompting-elicits-reasoning/). Advances in Neural Information Processing Systems, 35:24824–24837.
34. Niklas Muennighoff, Zitong Yang, Weijia Shi, Xiang Lisa Li, Li Fei-Fei, Hannaneh Hajishirzi, Luke Zettlemoyer, Percy Liang, Emmanuel Candès, and Tatsunori B Hashimoto. 2025b. [s1: Simple test-time scaling](https://md.research.elamur.ai/free/en/papers/muennighoff-2025-s1-simple-test-time-scaling/). In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, page…
35. Jianren Yang, Yikang Wang, Yang Zhao, Ziang Zhang, and Zhou Zhao. 2023. Leo: An embodied generalist agent in 3d world. arXiv preprint arXiv:2311.12871. — [arXiv:2311.12871](https://arxiv.org/abs/2311.12871) [find](https://scholar.google.com/scholar?q=Jianren+Yang%2C+Yikang+Wang%2C+Yang+Zhao%2C+Ziang+Zhang%2C+and+Zhou+Zhao%2E+2023%2E+Leo%3A+An+embodied+generalist+agent+in+3d+world%2E+arXiv+preprint+arXiv%3A2311%2E12871%2E+%E2%80%94+%5BarXiv%3A2311%2E12871%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F2311%2E)
36. OpenAI. 2023. [Gpt-4 technical report](https://md.research.elamur.ai/free/en/papers/openai-2023-gpt-4-technical-report/). openai.com/research/gpt-4. https://
37. OpenAI, Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, Red Avila, Igor Babuschkin, Suchir Balaji, Valerie Balcom, Paul Baltescu, Haiming Bao, Mohammad Bavarian, Jeff Belgum, and 262 others.… — [GPT-4 Technical Report](https://md.research.elamur.ai/free/en/papers/openai-2023-gpt-4-technical-report/)
38. Zhangyang Qi, Zhixiong Zhang, Ye Fang, Jiaqi Wang, and Hengshuang Zhao. 2025. Gpt4scene: Understand 3d scenes from videos with vision-language models. arXiv preprint arXiv:2501.01428. — [arXiv:2501.01428](https://arxiv.org/abs/2501.01428) [find](https://scholar.google.com/scholar?q=Zhangyang+Qi%2C+Zhixiong+Zhang%2C+Ye+Fang%2C+Jiaqi+Wang%2C+and+Hengshuang+Zhao%2E+2025%2E+Gpt4scene%3A+Understand+3d+scenes+from+videos+with+vision%2Dlanguage+models%2E+arXiv+preprint+arXiv%3A2501%2E01428%2E+%E2%80%94+%5BarXiv%3A2501%2E01)
39. Santhosh K Ramakrishnan, Aaron Gokaslan, Erik Wijmans, Oleksandr Maksymets, Alex Clegg, John Turner, Eric Undersander, Wojciech Galuba, Andrew Westbury, Angel X Chang, and 1 others. 2021. Habitat-matterport 3d dataset (hm3d): 1000 large-scale 3d environments for embodied ai. arXiv preprint arXiv:210… — [arXiv:2109.08238](https://arxiv.org/abs/2109.08238) [find](https://scholar.google.com/scholar?q=Santhosh+K+Ramakrishnan%2C+Aaron+Gokaslan%2C+Erik+Wijmans%2C+Oleksandr+Maksymets%2C+Alex+Clegg%2C+John+Turner%2C+Eric+Undersander%2C+Wojciech+Galuba%2C+Andrew+Westbury%2C+Angel+X+Chang%2C+and+1+others%2E+2021%2E+Habitat%2Dmatt)
40. Pranab Sahoo, Ayush Kumar Singh, Sriparna Saha, Vinija Jain, Samrat Mondal, and Aman Chadha. 2024. [A systematic survey of prompt engineering in large language models: Techniques and applications](https://md.research.elamur.ai/free/en/papers/sahoo-2024-systematic-survey-prompt-engineering-large/). arXiv preprint arXiv:2402.07927.
41. Charlie Snell, Jaehoon Lee, Kelvin Xu, and Aviral Kumar. 2024. [Scaling llm test-time compute optimally can be more effective than scaling model parameters](https://md.research.elamur.ai/free/en/papers/snell-2024-scaling-llm-test-time-compute/). arXiv preprint arXiv:2408.03314.
42. Zhou Yu, Jun Yu, Yuhao Cui, Dacheng Tao, and Qi Tian. 2019. Deep modular co-attention networks for visual question answering. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). [find](https://scholar.google.com/scholar?q=Zhou+Yu%2C+Jun+Yu%2C+Yuhao+Cui%2C+Dacheng+Tao%2C+and+Qi+Tian%2E+2019%2E+Deep+modular+co%2Dattention+networks+for+visual+question+answering%2E+In+Proceedings+of+the+IEEE%2FCVF+Conference+on+Computer+Vision+and+Pattern+R)
43. Sha Zhang, Di Huang, Jiajun Deng, Shixiang Tang, Wanli Ouyang, Tong He, and Yanyong Zhang. 2024a. Agent3d-zero: An agent for zero-shot 3d understanding. In European Conference on Computer Vision, pages 186–202. Springer. [find](https://scholar.google.com/scholar?q=Sha+Zhang%2C+Di+Huang%2C+Jiajun+Deng%2C+Shixiang+Tang%2C+Wanli+Ouyang%2C+Tong+He%2C+and+Yanyong+Zhang%2E+2024a%2E+Agent3d%2Dzero%3A+An+agent+for+zero%2Dshot+3d+understanding%2E+In+European+Conference+on+Computer+Vision%2C+page)
44. Yifan Zhang, Yikang Wang, Yang Zhao, Ziang Zhang, and Zhou Zhao. 2024b. Chatscene: Knowledge-enabled safety-critical scenario generation for autonomous vehicles. arXiv preprint arXiv:2405.14062. — [arXiv:2405.14062](https://arxiv.org/abs/2405.14062) [find](https://scholar.google.com/scholar?q=Yifan+Zhang%2C+Yikang+Wang%2C+Yang+Zhao%2C+Ziang+Zhang%2C+and+Zhou+Zhao%2E+2024b%2E+Chatscene%3A+Knowledge%2Denabled+safety%2Dcritical+scenario+generation+for+autonomous+vehicles%2E+arXiv+preprint+arXiv%3A2405%2E14062%2E+%E2%80%94+%5Bar)
45. Yuanhan Zhang, Jinming Wu, Wei Li, Bo Li, Zejun Ma, Ziwei Liu, and Chunyuan Li. 2024c. Video instruction tuning with synthetic data. Preprint, arXiv:2410.02713. [find](https://scholar.google.com/scholar?q=Yuanhan+Zhang%2C+Jinming+Wu%2C+Wei+Li%2C+Bo+Li%2C+Zejun+Ma%2C+Ziwei+Liu%2C+and+Chunyuan+Li%2E+2024c%2E+Video+instruction+tuning+with+synthetic+data%2E+Preprint%2C+arXiv%3A2410%2E02713%2E)
46. Duo Zheng, Shijia Huang, Lin Zhao, Yiwu Zhong, and Liwei Wang. 2024. Towards learning a generalist model for embodied navigation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). [find](https://scholar.google.com/scholar?q=Duo+Zheng%2C+Shijia+Huang%2C+Lin+Zhao%2C+Yiwu+Zhong%2C+and+Liwei+Wang%2E+2024%2E+Towards+learning+a+generalist+model+for+embodied+navigation%2E+In+Proceedings+of+the+IEEE%2FCVF+Conference+on+Computer+Vision+and+Patte)
47. Chenming Zhu, Tai Wang, Wenwei Zhang, Jiangmiao Pang, and Xihui Liu. 2024. Llava-3d: A simple yet effective pathway to empowering lmms with 3d-awareness. arXiv preprint arXiv:2409.18125. — [arXiv:2409.18125](https://arxiv.org/abs/2409.18125) [find](https://scholar.google.com/scholar?q=Chenming+Zhu%2C+Tai+Wang%2C+Wenwei+Zhang%2C+Jiangmiao+Pang%2C+and+Xihui+Liu%2E+2024%2E+Llava%2D3d%3A+A+simple+yet+effective+pathway+to+empowering+lmms+with+3d%2Dawareness%2E+arXiv+preprint+arXiv%3A2409%2E18125%2E+%E2%80%94+%5BarXiv%3A2409)
48. Ziyu Zhu, Xiaojian Ma, Yixin Chen, Zhidong Deng, Siyuan Huang, and Qing Li. 2023. 3d-vista: Pre-trained transformer for 3d vision and text alignment. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 2911– 2921. [find](https://scholar.google.com/scholar?q=Ziyu+Zhu%2C+Xiaojian+Ma%2C+Yixin+Chen%2C+Zhidong+Deng%2C+Siyuan+Huang%2C+and+Qing+Li%2E+2023%2E+3d%2Dvista%3A+Pre%2Dtrained+transformer+for+3d+vision+and+text+alignment%2E+In+Proceedings+of+the+IEEE%2FCVF+International+Confe)
49. Given a question and multiple viewpoints (including a bird-eye view), you must actively explore the scene by switching views and adjusting the camera to find the correct answer. You are REQUIRED to execute at least {{ min_action_steps }} action steps before answering. You are a helpful assistant tha… [find](https://scholar.google.com/scholar?q=Given+a+question+and+multiple+viewpoints+%28including+a+bird%2Deye+view%29%2C+you+must+actively+explore+the+scene+by+switching+views+and+adjusting+the+camera+to+find+the+correct+answer%2E+You+are+REQUIRED+to+ex)

*Updated 31 October 2026*