Pith. sign in

REVIEW 5 major objections 5 minor 1 cited by

MeshLLM: Empowering Large Language Models to Progressively Understand and Generate 3D Mesh

T0 review · 5 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read MeshLLM claims that decomposing 3D meshes into structurally meaningful primitive subunits lets large language models understand and generate text-serialized meshes better than prior methods, using a dataset roughly 50 times larger than…

desk verdict MeshLLM's real contributions could be solid, but the SOTA claim needs the full evaluation before it can be taken seriously. read the letter →

arxiv 2508.01242 v2 pith:EF2GQEPK submitted 2025-08-02 cs.GR cs.CV

Shuangkang Fang , I-Chao Shen , Yufeng Wang , Yi-Hsuan Tsai , Yi Yang , Shuchang Zhou , Wenrui Ding , Takeo Igarashi
show 1 more author
Ming-Hsuan Yang
This is my paper
classification cs.GRcs.CV
keywords 3Dmeshgenerationlargelanguagemodelstext-serializedmeshesprimitivedecompositionunderstandingtopologydatasetscaling
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

MeshLLM tries to establish that large language models can be made substantially better at both understanding and generating text-serialized 3D meshes by decomposing each mesh into structurally meaningful primitive subunits before serialization. This decomposition enables a training dataset of over 1.5 million samples, about 50 times larger than earlier datasets, which better matches the scaling behavior of large language models. The paper also introduces two training strategies: inferring face connectivity from vertex lists and local mesh assembly, both aimed at teaching the model real mesh topology and spatial structure. Reported experiments show MeshLLM outperforming the previous LLaMA-Mesh baseline in both mesh generation quality and shape understanding. A sympathetic reader cares because text-based 3D processing is a bottleneck for applying LLMs to geometry, and this is a concrete recipe for scaling past it.

What carries the argument

Primitive-Mesh decomposition is the central mechanism: it divides a 3D mesh into structurally meaningful subunits before text serialization, so each subpart is simpler and more regular for the LLM to process. This decomposition is what makes the 1.5 million-sample dataset possible, because it reduces the token length per sample and lets more meshes fit within LLM context windows. Two supporting mechanisms carry the topology learning: face connectivity inference, where the model is trained to recover faces from the vertex list of a decomposed part, and local mesh assembly, where the model learns to join parts back into a coherent whole. The combination of these mechanisms turns plain text-serialized meshes into a learnable 3D structure for the LLM.

What would settle it

One concrete test is to train a version of MeshLLM on the same 1.5 million samples but with the primitive decomposition applied randomly or inconsistently, for example cutting meshes at arbitrary planes rather than at structural boundaries, and then compare generation and understanding scores; if performance stays high, the gains are not due to structural decomposition. Alternatively, evaluating on meshes whose topology types are absent from the training set would show whether the connectivity training generalizes.

Watch

Extended reading notes

Core claim

The central claim is that the main obstacles to using LLMs for 3D meshes are dataset scale and the loss of structural information during serialization, and both can be repaired by a Primitive-Mesh decomposition that splits a mesh into meaningful subunits. On this decomposed representation, the paper builds a 1.5 million-sample dataset and adds two training signals: inferring face connectivity from the vertex sequence, and local mesh assembly. Together these let the model capture mesh topology and spatial layout rather than treating meshes as flat token strings. The paper reports that the resulting model surpasses LLaMA-Mesh in both generation quality and shape understanding, which supports the claim that LLMs can work directly on text-serialized geometry when the serialization is made structure-aware.

Load-bearing premise

The load-bearing premise is that breaking a mesh into primitive parts keeps enough geometric information for the model to learn true face connectivity and spatial assembly; if the decomposition discards relationships between parts, the reported gains will not hold beyond the training distribution.

Editorial extensions

If this is right

  • Text-serialized mesh processing by LLMs becomes scalable: the 1.5 million-sample corpus is roughly 50 times larger than previous datasets, matching the data-hungry nature of LLM training.
  • Mesh generation quality improves because the model is explicitly trained to infer face connectivity and assemble local parts, not just emit plausible token sequences.
  • Shape understanding improves with the same training signals, so a single model can handle both generation and comprehension of 3D meshes.
  • The approach ties mesh handling to the scaling-law paradigm: as more decomposed samples become available, LLM performance on 3D tasks should continue to improve.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The Primitive-Mesh decomposition strategy may generalize beyond meshes to other structured 3D representations, such as point clouds or parametric CAD models, where similar part-level decomposition could reduce token complexity.
  • A testable extension would be to ablate the face-connectivity and local-assembly objectives separately to measure which contributes more of the reported gain, since the abstract does not isolate their individual effects.
  • Because the dataset is built from primitives, the ceiling of the method may be set by the diversity of the primitive vocabulary; meshes whose structure is not expressible as combinations of the defined primitives could be underrepresented.
  • The gains may transfer to other LLM backbones and context lengths, suggesting the main barrier is representation and data rather than model architecture.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 5 minor

Summary. The paper presents MeshLLM, a framework for using large language models to understand and generate 3D meshes serialized as text. It introduces a Primitive-Mesh decomposition strategy that divides meshes into structurally meaningful subunits, enabling construction of a 1500k+ sample dataset approximately 50 times larger than prior work. The paper also proposes training strategies for inferring face connectivity from vertices and for local mesh assembly. The abstract reports that MeshLLM outperforms the state-of-the-art LLaMA-Mesh in both mesh generation quality and shape understanding. The manuscript as provided to the referee contains only this abstract; no methodology, experimental details, or quantitative comparisons are available for inspection.

Significance. If the stated results hold, MeshLLM would represent a meaningful advance in LLM-based 3D understanding and generation, particularly in addressing dataset scale and the loss of structural information during serialization. The decomposition strategy and topology-focused training are potentially valuable contributions that could influence subsequent work on text-serialized 3D data. However, the abstract alone provides no numerical evidence, no evaluation protocol, and no architectural or training details, so the significance cannot currently be evaluated beyond plausibility. The paper does not appear to ship code, proofs, or reproducible experiment artifacts based on the abstract alone.

major comments (5)
  1. [Abstract] The central claim that MeshLLM 'outperforms the state-of-the-art LLaMA-Mesh' is not supported by any quantitative results in the submitted text; the abstract omits the metrics, datasets, and statistical testing used to establish this claim, leaving the headline result unverifiable.
  2. [Abstract] The claim that the 1500k+ sample dataset is 'almost 50 times larger than previous methods' lacks a comparison baseline and does not specify how the Primitive-Mesh decomposition relates to token-length constraints; without these details, the scaling-law alignment argument cannot be assessed.
  3. [Abstract] The evaluation protocol is undefined with respect to the decomposition: if both training and testing use Primitive-Mesh serialized inputs, the reported improvement over LLaMA-Mesh may reflect a task-simplifying input representation rather than a general capability; the paper must include experiments on held-out raw (non-decomposed) meshes or standard benchmarks with arbitrary topology to establish transfer.
  4. [Abstract] The specific contributions of 'inferring face connectivity from vertices' and 'local mesh assembly' are not separated from the effects of dataset scale and decomposition; without ablations, it is unclear which component drives the claimed gains.
  5. [Full text (absent)] The submitted manuscript contains only an abstract; no architecture details, dataset construction procedure, training hyperparameters, or experimental setup are provided, so no aspect of the work can be reproduced or checked from this submission.
minor comments (5)
  1. [Abstract] The abbreviation LLaMA-Mesh is used without a reference to the original work.
  2. [Abstract] The term 'Primitive-Mesh' should be defined at first use, as it is not a standard term in the 3D mesh literature.
  3. [Abstract] The dataset size '1500k+' would be clearer as '1.5M' for readability.
  4. [Abstract] The phrase 'local mesh assembly' is not explained; a brief clarification would help a reader unfamiliar with the method.
  5. [Abstract] The abstract mentions alignment with 'LLM scaling law principles' but does not cite the scaling laws literature, making the connection vague.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found in the abstract; claims are empirical and no fitted parameters, self-citations, or definitional reductions are present.

full rationale

The abstract makes empirical claims about a new framework (MeshLLM) and reports that it outperforms LLaMA-Mesh on mesh generation quality and shape understanding. No equations, fitted parameters, or derivations are shown, so there is no visible step that reduces a predicted quantity to an input by construction. The Primitive-Mesh decomposition is introduced as a representation strategy, and the claimed benefit is an empirical outcome rather than an analytic equivalence. The dataset scale and benchmark comparisons are stated as results, not as consequences of a self-referential definition. No uniqueness theorems, self-citations, or ansatz-smuggling citations appear in the provided text. The lack of a detailed evaluation protocol could raise a correctness or generalization concern, but that is not circularity under the stated criteria. Since no specific circular step can be quoted and exhibited, the appropriate finding is no significant circularity.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

Only the abstract was reviewed, so the ledger reflects assumptions stated or implied there. No fitted parameters or invented entities were reported.

assumptions (3)
  • domain assumption Text-serialized meshes are an appropriate representation for LLM-based 3D understanding and generation.
    The whole framework depends on LLMs learning from serialized mesh tokens; this is asserted but not proven in the abstract.
  • ad hoc to paper Primitive-Mesh decomposition preserves the 3D structural information needed to infer face connectivity and assemble local pieces.
    This is a load-bearing modeling choice introduced by the paper; the abstract asserts it improves topology capture but provides no derivation.
  • domain assumption A dataset roughly 50 times larger than prior work aligns with LLM scaling law principles.
    The abstract invokes scaling laws to motivate the dataset size, but does not show a scaling curve or saturation analysis.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MeshLLM: Empowering Large Language Models to Progressively Understand and Generate 3D Mesh." pith.science (2026). https://pith.science/paper/EF2GQEPK

@misc{pith2026250801242,
  author       = {Pith},
  title        = {Pith review of: MeshLLM: Empowering Large Language Models to Progressively Understand and Generate 3D Mesh},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EF2GQEPK}},
  note         = {Machine review of arXiv:2508.01242}
}
read the original abstract

We present MeshLLM, a novel framework that leverages large language models (LLMs) to understand and generate text-serialized 3D meshes. Our approach addresses key limitations in existing methods, including the limited dataset scale when catering to LLMs' token length and the loss of 3D structural information during mesh serialization. We introduce a Primitive-Mesh decomposition strategy, which divides 3D meshes into structurally meaningful subunits. This enables the creation of a large-scale dataset with 1500k+ samples, almost 50 times larger than previous methods, which aligns better with the LLM scaling law principles. Furthermore, we propose inferring face connectivity from vertices and local mesh assembly training strategies, significantly enhancing the LLMs' ability to capture mesh topology and spatial structures. Experiments show that MeshLLM outperforms the state-of-the-art LLaMA-Mesh in both mesh generation quality and shape understanding, highlighting its great potential in processing text-serialized 3D meshes.

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. LL3M: Large Language 3D Modelers

    cs.GR 2025-08 conditional novelty 6.0 of 10

    A multi-agent LLM system generates editable 3D assets as Blender Python code, using documentation retrieval and visual self-critique to refine results.

Reference graph

Works this paper leans on

75 extracted references · 33 canonical work pages · cited by 1 Pith paper

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    Gpt-4 technical report

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical report. arXiv preprint arXiv:2303.08774, 2023

    arXiv 2023

  3. [3]

    Polydiff: Generating 3d polygonal meshes with diffusion models

    Antonio Alliegro, Yawar Siddiqui, Tatiana Tommasi, and Matthias Nie ner. Polydiff: Generating 3d polygonal meshes with diffusion models. arXiv preprint arXiv:2312.11417, 2023

    arXiv 2023

  4. [4]

    Palm 2 technical report

    Rohan Anil, Andrew M Dai, Orhan Firat, Melvin Johnson, Dmitry Lepikhin, Alexandre Passos, Siamak Shakeri, Emanuel Taropa, Paige Bailey, Zhifeng Chen, et al. Palm 2 technical report. arXiv preprint arXiv:2305.10403, 2023

    arXiv 2023

  5. [5]

    Qwen technical report

    Jinze Bai, Shuai Bai, Yunfei Chu, Zeyu Cui, Kai Dang, Xiaodong Deng, Yang Fan, Wenbin Ge, Yu Han, Fei Huang, et al. Qwen technical report. arXiv preprint arXiv:2309.16609, 2023

    arXiv 2023

  6. [6]

    Deepseek llm: Scaling open-source language models with longtermism

    Xiao Bi, Deli Chen, Guanting Chen, Shanhuang Chen, Damai Dai, Chengqi Deng, Honghui Ding, Kai Dong, Qiushi Du, Zhe Fu, et al. Deepseek llm: Scaling open-source language models with longtermism. arXiv preprint arXiv:2401.02954, 2024

    arXiv 2024

  7. [7]

    Language models are few-shot learners

    Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Language models are few-shot learners. Advances in Neural Information Processing Systems, 33: 0 1877--1901, 2020

    1901

  8. [8]

    Shapenet: An information-rich 3d model repository

    Angel X Chang, Thomas Funkhouser, Leonidas Guibas, Pat Hanrahan, Qixing Huang, Zimo Li, Silvio Savarese, Manolis Savva, Shuran Song, Hao Su, et al. Shapenet: An information-rich 3d model repository. arXiv preprint arXiv:1512.03012, 2015

    arXiv 2015

Show all 75 references
  1. [9]

    Meshxl: Neural coordinate field for generative 3d foundation models

    Sijin Chen, Xin Chen, Anqi Pang, Xianfang Zeng, Wei Cheng, Yijun Fu, Fukun Yin, Billzb Wang, Jingyi Yu, Gang Yu, et al. Meshxl: Neural coordinate field for generative 3d foundation models. Advances in Neural Information Processing Systems, 37: 0 97141--97166, 2025

    2025

  2. [10]

    Meshanything: Artist-created mesh generation with autoregressive transformers

    Yiwen Chen, Tong He, Di Huang, Weicai Ye, Sijin Chen, Jiaxiang Tang, Xin Chen, Zhongang Cai, Lei Yang, Gang Yu, et al. Meshanything: Artist-created mesh generation with autoregressive transformers. arXiv preprint arXiv:2406.10163, 2024 a

    2024 arXiv

  3. [11]

    Meshanything v2: Artist-created mesh generation with adjacent mesh tokenization

    Yiwen Chen, Yikai Wang, Yihao Luo, Zhengyi Wang, Zilong Chen, Jun Zhu, Chi Zhang, and Guosheng Lin. Meshanything v2: Artist-created mesh generation with adjacent mesh tokenization. arXiv preprint arXiv:2408.02555, 2024 b

    2024 arXiv

  4. [12]

    Grounded 3d-llm with referent tokens

    Yilun Chen, Shuai Yang, Haifeng Huang, Tai Wang, Runsen Xu, Ruiyuan Lyu, Dahua Lin, and Jiangmiao Pang. Grounded 3d-llm with referent tokens. arXiv preprint arXiv:2405.10370, 2024 c

    2024 arXiv

  5. [13]

    Sdfusion: Multimodal 3d shape completion, reconstruction, and generation

    Yen-Chi Cheng, Hsin-Ying Lee, Sergey Tulyakov, Alexander G Schwing, and Liang-Yan Gui. Sdfusion: Multimodal 3d shape completion, reconstruction, and generation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4456--4465, 2023

    2023

  6. [14]

    Palm: Scaling language modeling with pathways

    Aakanksha Chowdhery, Sharan Narang, Jacob Devlin, Maarten Bosma, Gaurav Mishra, Adam Roberts, Paul Barham, Hyung Won Chung, Charles Sutton, Sebastian Gehrmann, et al. Palm: Scaling language modeling with pathways. Journal of Machine Learning Research, 24 0 (240): 0 1--113, 2023

    2023

  7. [15]

    Scan2mesh: From unstructured range scans to 3d meshes

    Angela Dai and Matthias Nie ner. Scan2mesh: From unstructured range scans to 3d meshes. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5574--5583, 2019

    2019

  8. [16]

    Objaverse-xl: A universe of 10m+ 3d objects

    Matt Deitke, Ruoshi Liu, Matthew Wallingford, Huong Ngo, Oscar Michel, Aditya Kusupati, Alan Fan, Christian Laforte, Vikram Voleti, Samir Yitzhak Gadre, et al. Objaverse-xl: A universe of 10m+ 3d objects. Advances in Neural Information Processing Systems, 36: 0 35799--35813, 2023

    2023

  9. [17]

    Meteor universal: Language specific translation evaluation for any target language

    Michael Denkowski and Alon Lavie. Meteor universal: Language specific translation evaluation for any target language. In Proceedings of the Ninth Workshop on Statistical Machine Translation, pages 376--380, 2014

    2014

  10. [18]

    Enhancing chat language models by scaling high-quality instructional conversations

    Ning Ding, Yulin Chen, Bokai Xu, Yujia Qin, Zhi Zheng, Shengding Hu, Zhiyuan Liu, Maosong Sun, and Bowen Zhou. Enhancing chat language models by scaling high-quality instructional conversations. arXiv preprint arXiv:2305.14233, 2023

    2023 arXiv

  11. [19]

    Hyperdiffusion: Generating implicit neural fields with weight-space diffusion

    Ziya Erko c , Fangchang Ma, Qi Shan, Matthias Nie ner, and Angela Dai. Hyperdiffusion: Generating implicit neural fields with weight-space diffusion. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 14300--14310, 2023

    2023

  12. [20]

    Efficient implicit sdf and color reconstruction via shared feature field

    Shuangkang Fang, Dacheng Qi, Weixin Xu, Yufeng Wang, Zehao Zhang, Xiaorong Zhang, Huayu Zhang, Zeqi Shao, and Wenrui Ding. Efficient implicit sdf and color reconstruction via shared feature field. In Proceedings of the Asian Conference on Computer Vision, pages 3499--3516, 2024 a

    2024

  13. [21]

    Chat-edit-3d: Interactive 3d scene editing via text prompts

    Shuangkang Fang, Yufeng Wang, Yi-Hsuan Tsai, Yi Yang, Wenrui Ding, Shuchang Zhou, and Ming-Hsuan Yang. Chat-edit-3d: Interactive 3d scene editing via text prompts. In European Conference on Computer Vision, pages 199--216. Springer, 2024 b

    2024

  14. [22]

    Scene-llm: Extending language model for 3d visual understanding and reasoning

    Rao Fu, Jingyu Liu, Xilun Chen, Yixin Nie, and Wenhan Xiong. Scene-llm: Extending language model for 3d visual understanding and reasoning. arXiv preprint arXiv:2403.11401, 2024

    2024 arXiv

  15. [23]

    Worldgpt: Empowering llm as multimodal world model

    Zhiqi Ge, Hongzhe Huang, Mingze Zhou, Juncheng Li, Guoming Wang, Siliang Tang, and Yueting Zhuang. Worldgpt: Empowering llm as multimodal world model. In Proceedings of the 32nd ACM International Conference on Multimedia, pages 7346--7355, 2024

    2024

  16. [24]

    Imagebind: One embedding space to bind them all

    Rohit Girdhar, Alaaeldin El-Nouby, Zhuang Liu, Mannat Singh, Kalyan Vasudev Alwala, Armand Joulin, and Ishan Misra. Imagebind: One embedding space to bind them all. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 15180--15190, 2023

    2023

  17. [25]

    The llama 3 herd of models

    Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, et al. The llama 3 herd of models. arXiv e-prints, pages arXiv--2407, 2024

    2024

  18. [26]

    Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning

    Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, et al. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv preprint arXiv:2501.12948, 2025

    2025 arXiv

  19. [27]

    Point-bind & point-llm: Aligning point cloud with multi-modality for 3d understanding, generation, and instruction following

    Ziyu Guo, Renrui Zhang, Xiangyang Zhu, Yiwen Tang, Xianzheng Ma, Jiaming Han, Kexin Chen, Peng Gao, Xianzhi Li, Hongsheng Li, et al. Point-bind & point-llm: Aligning point cloud with multi-modality for 3d understanding, generation, and instruction following. arXiv preprint arX...

    2023 arXiv

  20. [28]

    3dgen: Triplane latent diffusion for textured mesh generation

    Anchit Gupta, Wenhan Xiong, Yixin Nie, Ian Jones, and Barlas O g uz. 3dgen: Triplane latent diffusion for textured mesh generation. arXiv preprint arXiv:2303.05371, 2023

    2023 arXiv

  21. [29]

    Audioclip: Extending clip to image, text and audio

    Andrey Guzhov, Federico Raue, J \"o rn Hees, and Andreas Dengel. Audioclip: Extending clip to image, text and audio. In IEEE International Conference on Acoustics, Speech and Signal Processing, pages 976--980. IEEE, 2022

    2022

  22. [30]

    Y2seq2seq: Cross-modal representation learning for 3d shape and text by joint reconstruction and prediction of view and word sequences

    Zhizhong Han, Mingyang Shang, Xiyang Wang, Yu-Shen Liu, and Matthias Zwicker. Y2seq2seq: Cross-modal representation learning for 3d shape and text by joint reconstruction and prediction of view and word sequences. In Proceedings of the AAAI Conference on Artificial Intelligenc...

    2019

  23. [31]

    Shapecaptioner: Generative caption network for 3d shapes by learning a mapping from parts detected in multiple views to sentences

    Zhizhong Han, Chao Chen, Yu-Shen Liu, and Matthias Zwicker. Shapecaptioner: Generative caption network for 3d shapes by learning a mapping from parts detected in multiple views to sentences. In Proceedings of the ACM International Conference on Multimedia, pages 1018--1027, 2020

    2020

  24. [32]

    Meshtron: High-fidelity, artist-like 3d mesh generation at scale

    Zekun Hao, David W Romero, Tsung-Yi Lin, and Ming-Yu Liu. Meshtron: High-fidelity, artist-like 3d mesh generation at scale. arXiv preprint arXiv:2412.09548, 2024

    2024 arXiv

  25. [33]

    Training compute-optimal large language models

    Jordan Hoffmann, Sebastian Borgeaud, Arthur Mensch, Elena Buchatskaya, Trevor Cai, Eliza Rutherford, Diego de Las Casas, Lisa Anne Hendricks, Johannes Welbl, Aidan Clark, et al. Training compute-optimal large language models. arXiv preprint arXiv:2203.15556, 2022

    2022 arXiv

  26. [34]

    3d-llm: Injecting the 3d world into large language models

    Yining Hong, Haoyu Zhen, Peihao Chen, Shuhong Zheng, Yilun Du, Zhenfang Chen, and Chuang Gan. 3d-llm: Injecting the 3d world into large language models. Advances in Neural Information Processing Systems, 36: 0 20482--20494, 2023

    2023

  27. [35]

    2d gaussian splatting for geometrically accurate radiance fields

    Binbin Huang, Zehao Yu, Anpei Chen, Andreas Geiger, and Shenghua Gao. 2d gaussian splatting for geometrically accurate radiance fields. In ACM SIGGRAPH 2024 Conference Papers, pages 1--11, 2024

    2024

  28. [36]

    Shap-e: Generating conditional 3d implicit functions

    Heewoo Jun and Alex Nichol. Shap-e: Generating conditional 3d implicit functions. arXiv preprint arXiv:2305.02463, 2023

    2023 arXiv

  29. [37]

    Scaling laws for neural language models

    Jared Kaplan, Sam McCandlish, Tom Henighan, Tom B Brown, Benjamin Chess, Rewon Child, Scott Gray, Alec Radford, Jeffrey Wu, and Dario Amodei. Scaling laws for neural language models. arXiv preprint arXiv:2001.08361, 2020

    2001 arXiv

  30. [38]

    3d gaussian splatting for real-time radiance field rendering

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimk \"u hler, and George Drettakis. 3d gaussian splatting for real-time radiance field rendering. ACM Trans. Graph., 42 0 (4): 0 139--1, 2023

    2023

  31. [39]

    Audio-llm: Activating the capabilities of large language models to comprehend audio data

    Dongting Li, Chenchong Tang, and Han Liu. Audio-llm: Activating the capabilities of large language models to comprehend audio data. In International Symposium on Neural Networks, pages 133--142. Springer, 2024

    2024

  32. [40]

    Acoustic prompt tuning: Empowering large language models with audition capabilities

    Jinhua Liang, Xubo Liu, Wenwu Wang, Mark D Plumbley, Huy Phan, and Emmanouil Benetos. Acoustic prompt tuning: Empowering large language models with audition capabilities. IEEE Transactions on Audio, Speech and Language Processing, 2025

    2025

  33. [41]

    Rouge: A package for automatic evaluation of summaries

    Chin-Yew Lin. Rouge: A package for automatic evaluation of summaries. In Text Summarization Branches out, pages 74--81, 2004

    2004

  34. [42]

    Deepseek-v2: A strong, economical, and efficient mixture-of-experts language model

    Aixin Liu, Bei Feng, Bin Wang, Bingxuan Wang, Bo Liu, Chenggang Zhao, Chengqi Dengr, Chong Ruan, Damai Dai, Daya Guo, et al. Deepseek-v2: A strong, economical, and efficient mixture-of-experts language model. arXiv preprint arXiv:2405.04434, 2024 a

    2024 arXiv

  35. [43]

    Deepseek-v3 technical report

    Aixin Liu, Bei Feng, Bing Xue, Bingxuan Wang, Bochao Wu, Chengda Lu, Chenggang Zhao, Chengqi Deng, Chenyu Zhang, Chong Ruan, et al. Deepseek-v3 technical report. arXiv preprint arXiv:2412.19437, 2024 b

    2024 arXiv

  36. [44]

    Visual instruction tuning

    Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. Visual instruction tuning. Advances in Neural Information Processing Systems, 36: 0 34892--34916, 2023

    2023

  37. [45]

    Improved baselines with visual instruction tuning

    Haotian Liu, Chunyuan Li, Yuheng Li, and Yong Jae Lee. Improved baselines with visual instruction tuning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 26296--26306, 2024 c

    2024

  38. [46]

    Music understanding llama: Advancing text-to-music generation with question answering and captioning

    Shansong Liu, Atin Sakkeer Hussain, Chenshuo Sun, and Ying Shan. Music understanding llama: Advancing text-to-music generation with question answering and captioning. In IEEE International Conference on Acoustics, Speech and Signal Processing, pages 286--290. IEEE, 2024 d

    2024

  39. [47]

    Nerf: Representing scenes as neural radiance fields for view synthesis

    Ben Mildenhall, Pratul P Srinivasan, Matthew Tancik, Jonathan T Barron, Ravi Ramamoorthi, and Ren Ng. Nerf: Representing scenes as neural radiance fields for view synthesis. Communications of the ACM, 65 0 (1): 0 99--106, 2021

    2021

  40. [48]

    Diffrf: Rendering-guided 3d radiance field diffusion

    Norman M \"u ller, Yawar Siddiqui, Lorenzo Porzi, Samuel Rota Bulo, Peter Kontschieder, and Matthias Nie ner. Diffrf: Rendering-guided 3d radiance field diffusion. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4328--4338, 2023

    2023

  41. [49]

    Using an llm to help with code understanding

    Daye Nam, Andrew Macvean, Vincent Hellendoorn, Bogdan Vasilescu, and Brad Myers. Using an llm to help with code understanding. In Proceedings of the IEEE/ACM 46th International Conference on Software Engineering, pages 1--13, 2024

    2024

  42. [50]

    Polygen: An autoregressive generative model of 3d meshes

    Charlie Nash, Yaroslav Ganin, SM Ali Eslami, and Peter Battaglia. Polygen: An autoregressive generative model of 3d meshes. In International Conference on Machine Learning, pages 7220--7229. PMLR, 2020

    2020

  43. [51]

    Training language models to follow instructions with human feedback

    Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, et al. Training language models to follow instructions with human feedback. Advances in Neural Information Processing Systems, 35: 0 2...

    2022

  44. [52]

    Bleu: a method for automatic evaluation of machine translation

    Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. Bleu: a method for automatic evaluation of machine translation. In Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics, pages 311--318, 2002

    2002

  45. [53]

    Deepsdf: Learning continuous signed distance functions for shape representation

    Jeong Joon Park, Peter Florence, Julian Straub, Richard Newcombe, and Steven Lovegrove. Deepsdf: Learning continuous signed distance functions for shape representation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 165--174, 2019

    2019

  46. [54]

    Can large language models understand symbolic graphics programs? In The Thirteenth International Conference on Learning Representations, 2024

    Zeju Qiu, Weiyang Liu, Haiwen Feng, Zhen Liu, Tim Z Xiao, Katherine M Collins, Joshua B Tenenbaum, Adrian Weller, Michael J Black, and Bernhard Sch \"o lkopf. Can large language models understand symbolic graphics programs? In The Thirteenth International Conference on Learnin...

    2024

  47. [55]

    Language models are unsupervised multitask learners

    Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. Language models are unsupervised multitask learners. OpenAI blog, 1 0 (8): 0 9, 2019

    2019

  48. [56]

    Learning transferable visual models from natural language supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervision. In International Conference on Machine Learning, pa...

    2021

  49. [57]

    Improved aesthetic predictor

    Christoph Schuhmann. Improved aesthetic predictor. https://github.com/christophschuhmann/improved-aesthetic-predictor, 2022

    2022

  50. [58]

    Llasm: Large language and speech model

    Yu Shu, Siwei Dong, Guangyao Chen, Wenhao Huang, Ruihua Zhang, Daochen Shi, Qiqi Xiang, and Yemin Shi. Llasm: Large language and speech model. arXiv preprint arXiv:2308.15930, 2023

    2023 arXiv

  51. [59]

    Meshgpt: Generating triangle meshes with decoder-only transformers

    Yawar Siddiqui, Antonio Alliegro, Alexey Artemov, Tatiana Tommasi, Daniele Sirigatti, Vladislav Rosov, Angela Dai, and Matthias Nie ner. Meshgpt: Generating triangle meshes with decoder-only transformers. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern...

    2024

  52. [60]

    Ernie 3.0: Large-scale knowledge enhanced pre-training for language understanding and generation

    Yu Sun, Shuohuan Wang, Shikun Feng, Siyu Ding, Chao Pang, Junyuan Shang, Jiaxiang Liu, Xuyi Chen, Yanbin Zhao, Yuxiang Lu, et al. Ernie 3.0: Large-scale knowledge enhanced pre-training for language understanding and generation. arXiv preprint arXiv:2107.02137, 2021

    2021 arXiv

  53. [61]

    Gemini: a family of highly capable multimodal models

    Gemini Team, Rohan Anil, Sebastian Borgeaud, Yonghui Wu, Jean-Baptiste Alayrac, Jiahui Yu, Radu Soricut, Johan Schalkwyk, Andrew M Dai, Anja Hauth, et al. Gemini: a family of highly capable multimodal models. arXiv preprint arXiv:2312.11805, 2023

    2023 arXiv

  54. [62]

    Llama: Open and efficient foundation language models

    Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timoth \'e e Lacroix, Baptiste Rozi \`e re, Naman Goyal, Eric Hambro, Faisal Azhar, et al. Llama: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971, 2023 a

    2023 arXiv

  55. [63]

    Llama 2: Open foundation and fine-tuned chat models

    Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, et al. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288, 2023 b

    2023 arXiv

  56. [64]

    Cider: Consensus-based image description evaluation

    Ramakrishna Vedantam, C Lawrence Zitnick, and Devi Parikh. Cider: Consensus-based image description evaluation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 4566--4575, 2015

    2015

  57. [65]

    Llama-mesh: Unifying 3d mesh generation with language models

    Zhengyi Wang, Jonathan Lorraine, Yikai Wang, Hang Su, Jun Zhu, Sanja Fidler, and Xiaohui Zeng. Llama-mesh: Unifying 3d mesh generation with language models. arXiv preprint arXiv:2411.09595, 2024

    2024 arXiv

  58. [66]

    Pivotmesh: Generic 3d mesh generation via pivot vertices guidance

    Haohan Weng, Yikai Wang, Tong Zhang, CL Chen, and Jun Zhu. Pivotmesh: Generic 3d mesh generation via pivot vertices guidance. arXiv preprint arXiv:2405.16890, 2024 a

    2024 arXiv

  59. [67]

    Scaling mesh generation via compressive tokenization

    Haohan Weng, Zibo Zhao, Biwen Lei, Xianghui Yang, Jian Liu, Zeqiang Lai, Zhuo Chen, Yuhong Liu, Jie Jiang, Chunchao Guo, et al. Scaling mesh generation via compressive tokenization. arXiv preprint arXiv:2411.07025, 2024 b

    2024 arXiv

  60. [68]

    Bloom: A 176b-parameter open-access multilingual language model

    BigScience Workshop, Teven Le Scao, Angela Fan, Christopher Akiki, Ellie Pavlick, Suzana Ili \'c , Daniel Hesslow, Roman Castagn \'e , Alexandra Sasha Luccioni, Fran c ois Yvon, et al. Bloom: A 176b-parameter open-access multilingual language model. arXiv preprint arXiv:2211.0...

    2022 arXiv

  61. [69]

    Chat2svg: Vector graphics generation with large language models and image diffusion models

    Ronghuan Wu, Wanchao Su, and Jing Liao. Chat2svg: Vector graphics generation with large language models and image diffusion models. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 23690--23700, 2025

    2025

  62. [70]

    Pointllm: Empowering large language models to understand point clouds

    Runsen Xu, Xiaolong Wang, Tai Wang, Yilun Chen, Jiangmiao Pang, and Dahua Lin. Pointllm: Empowering large language models to understand point clouds. In European Conference on Computer Vision, pages 131--147. Springer, 2024

    2024

  63. [71]

    Sampart3d: Segment any part in 3d objects

    Yunhan Yang, Yukun Huang, Yuan-Chen Guo, Liangjun Lu, Xiaoyang Wu, Edmund Y Lam, Yan-Pei Cao, and Xihui Liu. Sampart3d: Segment any part in 3d objects. arXiv preprint arXiv:2411.07184, 2024

    2024 arXiv

  64. [72]

    Shapegpt: 3d shape generation with a unified multi-modal language model

    Fukun Yin, Xin Chen, Chi Zhang, Biao Jiang, Zibo Zhao, Jiayuan Fan, Gang Yu, Taihao Li, and Tao Chen. Shapegpt: 3d shape generation with a unified multi-modal language model. arXiv preprint arXiv:2311.17618, 2023

    2023 arXiv

  65. [73]

    Speechgpt: Empowering large language models with intrinsic cross-modal conversational abilities

    Dong Zhang, Shimin Li, Xin Zhang, Jun Zhan, Pengyu Wang, Yaqian Zhou, and Xipeng Qiu. Speechgpt: Empowering large language models with intrinsic cross-modal conversational abilities. arXiv preprint arXiv:2305.11000, 2023 a

    2023 arXiv

  66. [74]

    Video-llama: An instruction-tuned audio-visual language model for video understanding

    Hang Zhang, Xin Li, and Lidong Bing. Video-llama: An instruction-tuned audio-visual language model for video understanding. arXiv preprint arXiv:2306.02858, 2023 b

    2023 arXiv

  67. [75]

    Vgbench: Evaluating large language models on vector graphics understanding and generation

    Bocheng Zou, Mu Cai, Jianrui Zhang, and Yong Jae Lee. Vgbench: Evaluating large language models on vector graphics understanding and generation. arXiv preprint arXiv:2407.10972, 2024

    2024 arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.