---
title: "Vision-Language Models Mistake Head Orientation for Gaze"
description: "The design and analysis plan were pre-registered. 32 references. Among them: 0."
authors: ["Zory Zhang", "Pinyuan Feng", "Bingyang Wang", "Tianwei Zhao", "Suyang Yu", "Qingying Gao", "Hokin Deng", "Ziqiao Ma", "Yijiang Li", "Dezhi Luo"]
year: 2026
venue: "Findings of the Association for Computational Linguistics: ACL 2026"
source: "https://aclanthology.org/2026.findings-acl.504/"
doi: "10.18653/v1/2026.findings-acl.504"
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/zhang-2026-vision-language-models-mistake-head/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/zhang-2026-vision-language-models-mistake-head/)

[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/55/) / Vision-Language Models Mistake Head Orientation for Gaze Direction: Nonverbal Conversation Cues

# Vision-Language Models Mistake Head Orientation for Gaze Direction: Nonverbal Conversation Cues

**Authors:** Zory Zhang, [Pinyuan Feng](https://md.research.elamur.ai/free/en/authors/pinyuan-feng/), Bingyang Wang, Tianwei Zhao, Suyang Yu, Qingying Gao, Hokin Deng, Ziqiao Ma, Yijiang Li and Dezhi Luo · **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.504/) · [PDF](https://aclanthology.org/2026.findings-acl.504.pdf) · [DOI 10.18653/v1/2026.findings-acl.504](https://doi.org/10.18653/v1/2026.findings-acl.504) · [OpenAlex](https://openalex.org/W4417095781)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
This study asks whether Vision-Language Models (VLMs) can determine which object a person is looking at, a nonverbal cue that people use to resolve references in conversation. The authors photographed 1,360 real scenes. In each, a seated person faces a table holding between two and four objects and looks at one of them. They also controlled head orientation: the head pointed at the target, pointed at a distractor, or was left free. Models answered a multiple-choice question naming the objects. Each stimulus was shown 10 times, and results were compared with human responses. The design and analysis plan were pre-registered. A pilot with 111 VLMs guided the choice of models for the main study.

VLMs performed far worse than humans. The authors ruled out image resolution and object-naming ability as explanations. The main cause is that models judge gaze from where the head points rather than from the eyes. A finetuning experiment on a transformer-based vision model suggests the bias comes from training data, not architecture. This points to targeted data as a way to improve gaze-aware human–machine interaction.

## Key points
- The main study tested GPT-4o-2024-08-06, GPT-5.2, Qwen3-VL-30B-A3B-Instruct, InternLM-XComposer2-vl-7b and GLM-4.6V. Two reference methods were added: Moondream2, which has a dedicated gaze function, and the specialized gaze model GazeLLE.
- The tests sort each method into one of five categories, ranging from gaze-competent to head-only. All evaluated systems showed consistent head-orientation biases.
- Finetuning GazeLLE's setup produced the largest gain when head and gaze disagreed: accuracy rose from 15.65% to 34.15%. When head and gaze agreed, accuracy rose only modestly, from 62.82% to 70.00%.
- The pre-registration is on OSF (osf.io/tj2wn). Both the 900 pilot stimuli and the main stimuli are publicly released.
- Stated limitations:
  - The gazers are few: one actor in the main study and two actresses in the pilot.
  - There is no data on how often head and gaze directions diverge in real life.

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

## Institutions

- [Carnegie Mellon University](https://md.research.elamur.ai/free/en/universities/carnegie-mellon-university/) — University, Pittsburgh, United States · 598 works · 191,226 citations
- [University of Washington](https://md.research.elamur.ai/free/en/universities/university-of-washington/) — University, Seattle, United States · 402 works · 212,763 citations
- [Johns Hopkins University](https://md.research.elamur.ai/free/en/universities/johns-hopkins-university/) — University, Baltimore, United States · 247 works · 106,235 citations
- [Columbia University](https://md.research.elamur.ai/free/en/universities/columbia-university/) — University, New York, United States · 254 works · 56,747 citations
- [Brown University](https://md.research.elamur.ai/free/en/universities/brown-university/) — University, Providence, United States · 80 works · 26,036 citations
- [University of Michigan](https://md.research.elamur.ai/free/en/universities/university-of-michigan/) — University, Ann Arbor, United States · 86 works · 15,493 citations
- [University of California San Diego](https://md.research.elamur.ai/free/en/universities/university-of-california-san-diego/) — University, San Diego, United States · 89 works · 54,032 citations
- [Emory University](https://md.research.elamur.ai/free/en/universities/emory-university/) — University, Atlanta, United States · 42 works · 39,899 citations

📚 Shelves (2): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [👁️ Computer Vision and Pattern Recognition](https://md.research.elamur.ai/free/en/shelves/computer-science/computer-vision-and-pattern-recognition/) · 3,486 works

🧭 Topics (2): [🧠 Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/) · 3,894 works · [🖼️ Vision & language](https://md.research.elamur.ai/free/en/topics/multimodal/) · 811 works

## Cite

Cite as: Zory Zhang, Pinyuan Feng, Bingyang Wang, Tianwei Zhao, Suyang Yu, Qingying Gao, Hokin Deng, Ziqiao Ma, Yijiang Li and Dezhi Luo. 2026. Vision-Language Models Mistake Head Orientation for Gaze Direction: Nonverbal Conversation Cues. 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{zhang2026vision,
  title = {{Vision-Language Models Mistake Head Orientation for Gaze Direction: Nonverbal Conversation Cues}},
  author = {Zory Zhang, Pinyuan Feng, Bingyang Wang, Tianwei Zhao, Suyang Yu, Qingying Gao, Hokin Deng, Ziqiao Ma, Yijiang Li and Dezhi Luo},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.504},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.504}
}
```

- [← previous VishBox v2: A Multi-Agent System for Adaptive Voice Phishing Simulation](https://md.research.elamur.ai/free/en/papers/park-2026-vishbox-v2-multi-agent-system/)
- [next → Visual and Memory–Augmented Soccer Commentary Generation](https://md.research.elamur.ai/free/en/papers/sun-2026-visual-memory-augmented-soccer-commentary/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/55/)

## References

32 references, 9 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. 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/). arXiv prepri…
2. Fazl Barez, Tung-Yu Wu, Iván Arcuschin, Michael Lan, Vincent Wang, Noah Siegel, Nicolas Collignon, Clement Neo, Isabelle Lee, Alasdair Paren, Adel Bibi, Robert Trager, Damiano Fornasiere, John Yan, Yanai Elazar, and Yoshua Bengio. 2025. Chain-of-thought is not explainability. (Back to Section 1) [find](https://scholar.google.com/scholar?q=Fazl+Barez%2C+Tung%2DYu+Wu%2C+Iv%C3%A1n+Arcuschin%2C+Michael+Lan%2C+Vincent+Wang%2C+Noah+Siegel%2C+Nicolas+Collignon%2C+Clement+Neo%2C+Isabelle+Lee%2C+Alasdair+Paren%2C+Adel+Bibi%2C+Robert+Trager%2C+Damiano+Fornasiere%2C+John+Yan%2C+Ya)
3. Douglas Bates, Martin Mächler, Ben Bolker, and Steve Walker. 2015. [Fitting linear mixed-effects models using lme4](https://md.research.elamur.ai/free/en/papers/bates-2015-fitting-linear-mixed-effects-models/). Journal of Statistical Software, 67(1):1–48. (Back to Section A.7)
4. Joshua R. de Leeuw, Rebecca A. Gilbert, and Björn Luchterhandt. 2023. jspsych: Enabling an open-source collaborative ecosystem of behavioral experiments. Journal of Open Source Software, 8(85):5351. (Back to Section A.4) [find](https://scholar.google.com/scholar?q=Joshua+R%2E+de+Leeuw%2C+Rebecca+A%2E+Gilbert%2C+and+Bj%C3%B6rn+Luchterhandt%2E+2023%2E+jspsych%3A+Enabling+an+open%2Dsource+collaborative+ecosystem+of+behavioral+experiments%2E+Journal+of+Open+Source+Software%2C+8%2885%29%3A5351%2E+%28)
5. Xiaoyi Dong, Pan Zhang, Yuhang Zang, Yuhang Cao, Bin Wang, Linke Ouyang, Xilin Wei, Songyang Zhang, Haodong Duan, Maosong Cao, Wenwei Zhang, Yining Li, Hang Yan, Yang Gao, Xinyue Zhang, Wei Li, Jingwen Li, Kai Chen, Conghui He, and 4 others. 2024. [Internlm-xcomposer2: Mastering free-form text-image…](https://md.research.elamur.ai/free/en/papers/dong-2024-internlm-xcomposer2-mastering-free/)
6. Haodong Duan, Junming Yang, Yuxuan Qiao, Xinyu Fang, Lin Chen, Yuan Liu, Xiaoyi Dong, Yuhang Zang, Pan Zhang, Jiaqi Wang, and 1 others. 2024. Vlmevalkit: An open-source toolkit for evaluating large multi-modality models. In Proceedings of the 32nd ACM International Conference on Multimedia, pages 11… [find](https://scholar.google.com/scholar?q=Haodong+Duan%2C+Junming+Yang%2C+Yuxuan+Qiao%2C+Xinyu+Fang%2C+Lin+Chen%2C+Yuan+Liu%2C+Xiaoyi+Dong%2C+Yuhang+Zang%2C+Pan+Zhang%2C+Jiaqi+Wang%2C+and+1+others%2E+2024%2E+Vlmevalkit%3A+An+open%2Dsource+toolkit+for+evaluating+large+mu)
7. Yotam Erel, Katherine Adams Shannon, Junyi Chu, Kim Scott, Melissa Kline Struhl, Peng Cao, Xincheng Tan, Peter Hart, Gal Raz, Sabrina Piccolo, Catherine Mei, Christine Potter, Sagi Jaffe-Dax, Casey Lew-Williams, Joshua Tenenbaum, Katherine Fairchild, Amit Bermano, and Shari Liu. 2023. icatcher+: Rob… [find](https://scholar.google.com/scholar?q=Yotam+Erel%2C+Katherine+Adams+Shannon%2C+Junyi+Chu%2C+Kim+Scott%2C+Melissa+Kline+Struhl%2C+Peng+Cao%2C+Xincheng+Tan%2C+Peter+Hart%2C+Gal+Raz%2C+Sabrina+Piccolo%2C+Catherine+Mei%2C+Christine+Potter%2C+Sagi+Jaffe%2DDax%2C+Casey+Le)
8. Roberta Michnick Golinkoff, Kathryn Hirsh-Pasek, Kathleen M. Cauley, and Laura Gordon. 1987. The eyes have it: lexical and syntactic comprehension in a new paradigm. Journal of Child Language, 14(1):23–45. (Back to Section 1) [find](https://scholar.google.com/scholar?q=Roberta+Michnick+Golinkoff%2C+Kathryn+Hirsh%2DPasek%2C+Kathleen+M%2E+Cauley%2C+and+Laura+Gordon%2E+1987%2E+The+eyes+have+it%3A+lexical+and+syntactic+comprehension+in+a+new+paradigm%2E+Journal+of+Child+Language%2C+14%281%29%3A2)
9. Anshul Gupta, Pierre Vuillecard, Arya Farkhondeh, and Jean-Marc Odobez. 2024. Exploring the zero-shot capabilities of vision-language models for improving gaze following. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 615–624. (Back to Sections: 1, 2.2) [find](https://scholar.google.com/scholar?q=Anshul+Gupta%2C+Pierre+Vuillecard%2C+Arya+Farkhondeh%2C+and+Jean%2DMarc+Odobez%2E+2024%2E+Exploring+the+zero%2Dshot+capabilities+of+vision%2Dlanguage+models+for+improving+gaze+following%2E+In+Proceedings+of+the+IEEE%2FCV)
10. Nicole X. Han, William Yang Wang, and Miguel P. Eckstein. 2021. Gaze perception in humans and cnn-based model. Preprint, arXiv:2104.08447. (Back to Sections: 1, 2.2) — [arXiv:2104.08447](https://arxiv.org/abs/2104.08447) [find](https://scholar.google.com/scholar?q=Nicole+X%2E+Han%2C+William+Yang+Wang%2C+and+Miguel+P%2E+Eckstein%2E+2021%2E+Gaze+perception+in+humans+and+cnn%2Dbased+model%2E+Preprint%2C+arXiv%3A2104%2E08447%2E+%28Back+to+Sections%3A+1%2C+2%2E2%29+%E2%80%94+%5BarXiv%3A2104%2E08447%5D%28https%3A%2F%2Farxiv)
11. Subbarao Kambhampati. 2024. Can large language models reason and plan? Annals of the New York Academy of Sciences, 1534(1):15–18. (Back to Section 1) [find](https://scholar.google.com/scholar?q=Subbarao+Kambhampati%2E+2024%2E+Can+large+language+models+reason+and+plan%3F+Annals+of+the+New+York+Academy+of+Sciences%2C+1534%281%29%3A15%E2%80%9318%2E+%28Back+to+Section+1%29)
12. Kobin H. Kendrick, Judith Holler, and Stephen C. Levinson. 2023. Turn-taking in human face-to-face interaction is multimodal: gaze direction and manual gestures aid the coordination of turn transitions. Philosophical Transactions of the Royal Society B: Biological Sciences, 378(1875):20210473. (Back… [find](https://scholar.google.com/scholar?q=Kobin+H%2E+Kendrick%2C+Judith+Holler%2C+and+Stephen+C%2E+Levinson%2E+2023%2E+Turn%2Dtaking+in+human+face%2Dto%2Dface+interaction+is+multimodal%3A+gaze+direction+and+manual+gestures+aid+the+coordination+of+turn+transition)
13. Michael F. Land. 2006. Eye movements and the control of actions in everyday life. Progress in Retinal and Eye Research, 25(3):296–324. (Back to Section 1) [find](https://scholar.google.com/scholar?q=Michael+F%2E+Land%2E+2006%2E+Eye+movements+and+the+control+of+actions+in+everyday+life%2E+Progress+in+Retinal+and+Eye+Research%2C+25%283%29%3A296%E2%80%93324%2E+%28Back+to+Section+1%29)
14. Drew Linsley, Pinyuan Feng, and Thomas Serre. 2025. Better artificial intelligence does not mean better models of biology. Trends in Cognitive Sciences. (Back to Section 6) [find](https://scholar.google.com/scholar?q=Drew+Linsley%2C+Pinyuan+Feng%2C+and+Thomas+Serre%2E+2025%2E+Better+artificial+intelligence+does+not+mean+better+models+of+biology%2E+Trends+in+Cognitive+Sciences%2E+%28Back+to+Section+6%29)
15. Drew Linsley, Peisen Zhou, Alekh Karkada Ashok, Akash Nagaraj, Gaurav Gaonkar, Francis E Lewis, Zygmunt Pizlo, and Thomas Serre. 2024. The 3d-pc: a benchmark for visual perspective taking in humans and machines. arXiv preprint. (Back to Section 6) [find](https://scholar.google.com/scholar?q=Drew+Linsley%2C+Peisen+Zhou%2C+Alekh+Karkada+Ashok%2C+Akash+Nagaraj%2C+Gaurav+Gaonkar%2C+Francis+E+Lewis%2C+Zygmunt+Pizlo%2C+and+Thomas+Serre%2E+2024%2E+The+3d%2Dpc%3A+a+benchmark+for+visual+perspective+taking+in+humans+an)
16. Changsong Liu, Rui Fang, and Joyce Y. Chai. 2013. Shared Gaze in Situated Referential Grounding: An Empirical Study, pages 23–39. Springer London, London. (Back to Section 1) [find](https://scholar.google.com/scholar?q=Changsong+Liu%2C+Rui+Fang%2C+and+Joyce+Y%2E+Chai%2E+2013%2E+Shared+Gaze+in+Situated+Referential+Grounding%3A+An+Empirical+Study%2C+pages+23%E2%80%9339%2E+Springer+London%2C+London%2E+%28Back+to+Section+1%29)
17. Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. 2023. [Visual instruction tuning](https://md.research.elamur.ai/free/en/papers/liu-2023-visual-instruction-tuning/). Preprint, arXiv:2304.08485. (Back to Section 1)
18. Ramesh Manuvinakurike, Emanuel Moss, Elizabeth Anne Watkins, Saurav Sahay, Giuseppe Raffa, and Lama Nachman. 2025. Thoughts without thinking: Reconsidering the explanatory value of chain-of-thought reasoning in llms through agentic pipelines. Preprint, arXiv:2505.00875. (Back to Section 1) — [arXiv:2505.00875](https://arxiv.org/abs/2505.00875) [find](https://scholar.google.com/scholar?q=Ramesh+Manuvinakurike%2C+Emanuel+Moss%2C+Elizabeth+Anne+Watkins%2C+Saurav+Sahay%2C+Giuseppe+Raffa%2C+and+Lama+Nachman%2E+2025%2E+Thoughts+without+thinking%3A+Reconsidering+the+explanatory+value+of+chain%2Dof%2Dthought+re)
19. Moondream.ai. 2025. Moondream: Announcing gaze detection. (Back to Section 2.1) [find](https://scholar.google.com/scholar?q=Moondream%2Eai%2E+2025%2E+Moondream%3A+Announcing+gaze+detection%2E+%28Back+to+Section+2%2E1%29)
20. OpenAI. 2024. [Gpt-4o system card](https://md.research.elamur.ai/free/en/papers/openai-2024-gpt-4o-system-card/). (Back to Section A.1)
21. OpenAI. 2025. Gpt-5.2 system card. (Back to Sections: 2.1, A.1) [find](https://scholar.google.com/scholar?q=OpenAI%2E+2025%2E+Gpt%2D5%2E2+system+card%2E+%28Back+to+Sections%3A+2%2E1%2C+A%2E1%29)
22. Michael I. Posner and Steven E. Petersen. 1990. The attention system of the human brain. Annual Review of Neuroscience, 13(Volume 13, 1990):25–42. (Back to Section 1) [find](https://scholar.google.com/scholar?q=Michael+I%2E+Posner+and+Steven+E%2E+Petersen%2E+1990%2E+The+attention+system+of+the+human+brain%2E+Annual+Review+of+Neuroscience%2C+13%28Volume+13%2C+1990%29%3A25%E2%80%9342%2E+%28Back+to+Section+1%29)
23. Zahar Prasov and Joyce Y Chai. 2008. What’s in a gaze? the role of eye-gaze in reference resolution in multimodal conversational interfaces. In Proceedings of the 13th international conference on Intelligent user interfaces, pages 20–29. (Back to Section 1) [find](https://scholar.google.com/scholar?q=Zahar+Prasov+and+Joyce+Y+Chai%2E+2008%2E+What%E2%80%99s+in+a+gaze%3F+the+role+of+eye%2Dgaze+in+reference+resolution+in+multimodal+conversational+interfaces%2E+In+Proceedings+of+the+13th+international+conference+on+Inte)
24. Kun Qian, Zhuoyang Zhang, Wei Song, and Jianfeng Liao. 2023. Gvgnet: Gaze-directed visual grounding for learning under-specified object referring intention. IEEE Robotics and Automation Letters, 8(9):5990– 5997. (Back to Section 1) [find](https://scholar.google.com/scholar?q=Kun+Qian%2C+Zhuoyang+Zhang%2C+Wei+Song%2C+and+Jianfeng+Liao%2E+2023%2E+Gvgnet%3A+Gaze%2Ddirected+visual+grounding+for+learning+under%2Dspecified+object+referring+intention%2E+IEEE+Robotics+and+Automation+Letters%2C+8%289%29%3A)
25. Adria Recasens, Aditya Khosla, Carl Vondrick, and Antonio Torralba. 2015. Where are they looking? In Advances in Neural Information Processing Systems (NIPS). (Back to Section 3.1) [find](https://scholar.google.com/scholar?q=Adria+Recasens%2C+Aditya+Khosla%2C+Carl+Vondrick%2C+and+Antonio+Torralba%2E+2015%2E+Where+are+they+looking%3F+In+Advances+in+Neural+Information+Processing+Systems+%28NIPS%29%2E+%28Back+to+Section+3%2E1%29)
26. Alexander Robitzsch. 2020. Why ordinal variables can (almost) always be treated as continuous variables: Clarifying assumptions of robust continuous and ordinal factor analysis estimation methods. Frontiers in Education, Volume 5 - 2020. (Back to Section 3) [find](https://scholar.google.com/scholar?q=Alexander+Robitzsch%2E+2020%2E+Why+ordinal+variables+can+%28almost%29+always+be+treated+as+continuous+variables%3A+Clarifying+assumptions+of+robust+continuous+and+ordinal+factor+analysis+estimation+methods%2E+Fro)
27. Fiona Ryan, Ajay Bati, Sangmin Lee, Daniel Bolya, Judy Hoffman, and James M. Rehg. 2025. Gaze-lle: Gaze target estimation via large-scale learned encoders. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. (Back to Sections: 1, 2.1, and 2.2) [find](https://scholar.google.com/scholar?q=Fiona+Ryan%2C+Ajay+Bati%2C+Sangmin+Lee%2C+Daniel+Bolya%2C+Judy+Hoffman%2C+and+James+M%2E+Rehg%2E+2025%2E+Gaze%2Dlle%3A+Gaze+target+estimation+via+large%2Dscale+learned+encoders%2E+In+Proceedings+of+the+IEEE%2FCVF+Conference+on)
28. 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/)
29. Team Gemini. 2023. [Gemini: A family of highly capable multimodal models](https://md.research.elamur.ai/free/en/papers/gemini-team-2023-gemini-family-highly-capable-multimodal/). Preprint, arXiv:2312.11805. (Back to Section 1)
30. Team Gemini. 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/). Preprint, arXiv:2403.05530. (Back to Section A.1)
31. Team GLM and 1 others. 2024. [Chatglm: A family of large language models from glm-130b to glm-4 all tools](https://md.research.elamur.ai/free/en/papers/glm-2024-chatglm-family-large-language-models/). Preprint, arXiv:2406.12793. (Back to Section A.1)
32. Team Qwen. 2025. Qwen2.5-vl. (Back to Section A.1) [find](https://scholar.google.com/scholar?q=Team+Qwen%2E+2025%2E+Qwen2%2E5%2Dvl%2E+%28Back+to+Section+A%2E1%29)

*Updated 2 November 2026*