Pith. sign in

REVIEW 3 major objections 3 minor 42 references

Evaluating Model Perception of Color Illusions in Photorealistic Scenes

T0 review · 3 major / 3 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Current vision-language models perceive color illusions much as humans do, a new 19,000-image photorealistic dataset shows—and they can be trained to report both human-like and pixel-based judgments.

desk verdict RCID is a genuinely useful dataset, but the pixel-preservation assumption is load-bearing and unverified, so the human-similarity claim needs softening. read the letter →

arxiv 2412.06184 v1 pith:OBYCAJZH submitted 2024-12-09 cs.CV

Lingjun Mao , Zineng Tang , Alane Suhr This is my paper · ORCID
classification cs.CV
keywords colorillusionsvision-languagemodelsperceptualbiasphotorealisticdatasetcontrastillusionstripefilterinstructiontuning
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

The paper asks whether vision-language models (VLMs) are susceptible to color illusions the way humans are, and proposes a way to test this at scale. It introduces an automated pipeline that turns simple, procedurally generated illusion images into photorealistic scenes, producing RCID, a dataset of 19,000 images spanning contrast, stripe, and filter illusions, half of which contain an illusion and half of which are controls. Each image is paired with two questions: one asking about color according to pixel values, the other according to human perception. After fine-tuning on non-illusion images, all studied VLMs score high on non-illusion images but drop markedly on illusion images, and a substantial share of their wrong answers match human perception rather than pixel values. The paper concludes that current VLMs inherit human-like perceptual biases for color illusions, and that a simple mixed-training baseline can make a model report either human-perceived or pixel-based colors on request.

What carries the argument

The mechanism that carries the argument is a three-stage dataset-generation pipeline combined with paired-question evaluation. Simple illusion images are first generated procedurally (backgrounds with different brightness, stripes with alternating colors, etc.), then a conditional diffusion model maps them to photorealistic scenes while retaining the intended color relationships; filter illusions are made by shifting hue in HSV space on MS COCO images. For every image, GPT-4o generates a natural-language question asking about color differences or object color, and two prompts are attached: one asks for the judgment 'based on pixel values' and the other 'based on human perception.' Human validation on Prolific decides which images truly deceive at least 3 of 5 viewers. This design lets the paper label each image with two ground truths—pixel truth and human truth—so a model response can be classified as accurate, human-like, or neither.

What would settle it

Compare, on a random sample of RCID contrast and stripe images, the mean RGB values of the two regions named in the question. If a nontrivial fraction of images labeled 'same' actually have different mean pixel colors (or vice versa), the pixel-based ground truth is broken and the reported deception rates would need to be recomputed. One could also re-run the VLM evaluation on the procedural source images versus the ControlNet-realistic versions: if the human-like response rate changes substantially, the photorealistic rendering step is introducing confounds.

Watch

Extended reading notes

Core claim

The central claim is that current vision-language models exhibit perceptual biases similar to human vision when confronted with realistic color illusions. The evidence comes from RCID: models fine-tuned to answer color-comparison questions reach over 75% accuracy on non-illusion images, while accuracy on illusion images is significantly lower, and a portion of the errors are 'human-like'—they agree with human judgments rather than with pixel values. The paper also reports that model size increases human-like responding, that external prompts such as chain-of-thought or few-shot examples shift responses only modestly, and that fine-tuning on illusion images can push models toward either pixel-based or human-like answers. A mixed-training baseline lets a single model produce the correct pixel-based answer under a 'pixel values' prompt and the correct human-based answer under a 'human perception' prompt.

Load-bearing premise

The photorealistic images produced by the conditional diffusion model preserve the exact pixel-level color relationships of the procedurally generated source images, so that the 'based on pixel values' ground-truth labels are correct for every illusion image.

Editorial extensions

If this is right

  • Because VLMs show human-like biases on photorealistic illusions, downstream applications that rely on color judgments in real images will inherit those systematic errors.
  • The success of mixed instruction tuning shows that models can be trained to report pixel-based or human-perception answers on demand, giving application developers a lever to choose the appropriate behavior.
  • The small effect of chain-of-thought and few-shot prompting indicates that these biases are not superficial output artifacts but are rooted in visual representations and learned priors.
  • The positive correlation between model size and human-like responses suggests that scaling model size alone will not eliminate these perceptual biases.

Reading between the lines

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

  • If the same pattern holds more broadly, color-illusion susceptibility could serve as a probe for how much a VLM's judgments rely on low-level visual processing versus stored world knowledge, complementing other perceptual benchmarks.
  • One testable extension is to vary the photorealistic generation process (e.g., different diffusion models or strengths of realism) and check whether human-like response rates track the fidelity of the scene, which would separate the contribution of image realism from the model's internal priors.
  • The RCID-style paired-question setup could be adapted to other perceptual constancies, such as lightness, size, or shape illusions, to see whether 'human-like' errors are a general property of VLM visual encoders or specific to color.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 3 minor

Summary. The paper proposes an automated pipeline for generating photorealistic color-illusion images and introduces RCID, a dataset of 19,000 images spanning contrast, stripe, and filter illusions, with human validation labels. The authors evaluate several open-source VLMs after fine-tuning on non-illusion images, reporting strong accuracy on non-illusion images but significantly lower accuracy on illusion images, with a substantial fraction of human-like responses. They also investigate the roles of the visual encoder and linguistic/commonsense priors, and propose a mixed-training method that lets a model answer according to either pixel values or human perception.

Significance. If the results hold, this is a valuable contribution: RCID is much larger and more realistic than prior illusion datasets, the human-annotation protocol is independent and the fine-tuning design helps separate color-comparison ability from illusion susceptibility. The paper also goes beyond benchmark reporting by studying model size, prompt effects, and the visual/linguistic sources of bias. The code and data are promised open-source, which strengthens reproducibility. The main risk is that the pixel-level ground truth of the generated images is never verified; this threatens the validity of the no-illusion/deception metrics and, in turn, the central claim of human-like perceptual bias.

major comments (3)
  1. [3.1] The generation stage as written is not a standard diffusion process and is inconsistent with the training objective. Equation (1) trains the model to denoise from x0 = G (the quantized image), but the generating-stage formula I' = sum_t (x_t + epsilon_theta(...)) sums noisy samples and predicted noise across all time steps. This is not the usual recursive denoising update and it is unclear what x_t denotes after the first step. Because the paper's central claim depends on the final images preserving the intended color relationships, the actual generation algorithm must be specified precisely and reproducibly.
  2. [3.1 and 4.1] The pixel-value ground truth used for No-Illusion and Deception rates in Section 4 rests on the assumption that the ControlNet-generated images I' preserve the mean-color relationships of the procedural images S (e.g., the two target squares have identical pixel colors in 'same but look different' contrast illusions, and differ in the intended direction in 'different but look same' cases). This is never verified after generation. Human validation in Section 3.4 only confirms that participants were deceived; it does not confirm that the final images satisfy the intended pixel comparisons. ControlNet is conditioned on text and noise, so local color shifts are plausible; if one target region shifts more than another, the reported >75% non-illusion accuracy and the accuracy drop on illusion images could partly be an artifact of mislabeled pixel comparisons rather than genuine perceptual bias. The authors should measure the target-region mean colors in the final images, filter or re-label images that fail the intended comparisons, and rerun the affected metrics.
  3. [Abstract and Appendix B.4] The abstract's claim that 'all studied VLMs exhibit perceptual biases similar human vision' is stronger than the evidence in the paper. Appendix B.4 shows that without fine-tuning, models perform near random guessing on contrast and stripe tasks even on non-illusion images, so their 'human-like' behavior on illusions is not distinguishable from chance at the base-model level. The human-like bias is convincingly demonstrated only for models after fine-tuning on non-illusion images. The claim should be qualified to the fine-tuned setting, or the base-model analysis should be used to support a weaker claim about susceptibility rather than similarity to human perception.
minor comments (3)
  1. [Abstract and 6] There are typos: 'similar human vision' should be 'similar to human vision' in the abstract; 'an simple training method' in Section 1; 'fileds' in Section 6.
  2. [3.1] The notation is inconsistent: x0 is defined as the quantized grid G in the training stage but as the procedural image S in the generating stage, and the symbol C is used both for colors and for text captions. Please use distinct symbols.
  3. [5.2] The analysis in Figure 10 reports correlations between color descriptor diversity and pixel-value accuracy, but the figure would benefit from error bars or a clearer statement of the number of color terms per language and the significance of the correlation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: RCID labels come from independent human annotators and procedural pixel definitions, and VLM scores are measured against those external labels.

full rationale

The paper's central comparison is not circular: human-perception ground truth is obtained from independent Prolific annotators (Section 3.4), pixel-value ground truth is defined by the procedural color assignments (Section 3.1), and VLM responses are then scored against these external labels (Section 4.1). The fine-tuning experiments are standard supervised learning with held-out test splits; the 'mixed training' model is trained on both label types and evaluated on a separate test set, so its ability to follow pixel versus perception prompts is not forced by construction. The main validity threat, that ControlNet may not preserve exact pixel equalities or differences in the generated photorealistic images, would make some external labels incorrect, but it is an empirical assumption about image fidelity rather than a derivation that reduces to the paper's own inputs. Self-citations (e.g., [9], [35]) are contextual and carry no load in the evaluation chain. Accordingly, no circular step is identified.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The dataset's validity rests on human labels for perceived color and on an unverified assumption that the photorealistic rendering preserves the pixel-value relationships of the simple input images. The generation parameters are hand-set and the labeling threshold (3 of 5) is a choice, both of which shape the dataset statistics but are not fitted to model outputs. No invented entities are introduced.

free parameters (4)
  • Contrast illusion foreground/background brightness scales (µb1, µb2, µf1, µf2) = randomly sampled ranges (not reported exactly)
    These scaling factors determine the strength of the generated contrast illusions and are chosen by the procedural generator, not fitted to model outputs. They could influence which images are classified as illusions, but the paper does not list the sampling ranges.
  • Stripe illusion color scales (µs1, µs2) = based on initial human-perception estimates
    Chosen by hand to create stripe illusions of intended strength; affects the dataset composition.
  • Human deception threshold (3 of 5 annotators) = 3/5
    An image is labeled as an illusion if at least three of five Prolific participants are deceived. Changing this threshold changes the dataset and the human-perception ground truth.
  • Fine-tuning epochs (5) and data mixture sizes = 5 epochs; 6000 simple + 5500 realistic non-illusion images
    Hyperparameters for the model fine-tuning; not fitted to the target result but affect the reported accuracy numbers.
assumptions (4)
  • domain assumption ControlNet training on MS COCO produces a mapping that can turn the simple procedural illusion images into photorealistic images while preserving the illusion and the relative pixel colors.
    This assumption underlies the whole dataset construction in Section 3.1. The paper provides no direct check that pixel relationships in the final images match the input simple images.
  • domain assumption The majority vote of five Prolific participants is a reliable estimate of human perception for each image.
    Used in Section 3.4 to define illusion vs non-illusion labels; with only five annotators the labels are noisy, though Fleiss' kappa reported afterward is moderate.
  • domain assumption GPT-4o-generated questions accurately and unambiguously refer to the intended image regions without leaking the correct answers.
    Section 3.3 relies on generated questions; the prompts avoid some keywords but biased or ambiguous questions could affect both human and model responses.
  • domain assumption The color illusions studied (contrast, stripe, filter) are representative of real-world color perception phenomena.
    The paper selects these three families; conclusions about 'VLMs and color illusions' are scoped to these types.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Evaluating Model Perception of Color Illusions in Photorealistic Scenes." pith.science (2026). https://pith.science/paper/OBYCAJZH

@misc{pith2026241206184,
  author       = {Pith},
  title        = {Pith review of: Evaluating Model Perception of Color Illusions in Photorealistic Scenes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OBYCAJZH}},
  note         = {Machine review of arXiv:2412.06184}
}
read the original abstract

We study the perception of color illusions by vision-language models. Color illusion, where a person's visual system perceives color differently from actual color, is well-studied in human vision. However, it remains underexplored whether vision-language models (VLMs), trained on large-scale human data, exhibit similar perceptual biases when confronted with such color illusions. We propose an automated framework for generating color illusion images, resulting in RCID (Realistic Color Illusion Dataset), a dataset of 19,000 realistic illusion images. Our experiments show that all studied VLMs exhibit perceptual biases similar human vision. Finally, we train a model to distinguish both human perception and actual pixel differences.

Figures

Figures reproduced from arXiv: 2412.06184 by the authors.

Figure 1
Figure 1. Examples of color illusions. (a) Contrast illusion: The two green squares on the left and right are identical in color, but most people perceive the square on the right as darker due to the brighter background. (b) Stripe illusion: The left and right walls seem to be different colors, but this is an illusion caused by the dark stripes interfering with our perception. (c) Filter illusion: The player’s uniform appears… view at source ↗
Figure 2
Figure 2. Process for generating our dataset. 3.1. Contrast and Stripe Illusion Image Generation ControlNet Training Stage: We begin by training a Con￾trolNet to map from simple illusion images to photoreal￾istic images with illusions. We train using modified image￾caption pairs (I, T) from the MS COCO 2017 dataset. First, we quantize each original image I into a 10 × 10 grid G, where each cell Gx,y represents the average col… view at source ↗
Figure 3
Figure 3. Data statistics of RCID (Realistic Color Illusion Dataset). [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (15 more)
Figure 4
Figure 4. Figure 4: This figure shows the proportion of different model responses across three types of illusions (Contrast, Filter, and Stripe) on [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Deception rates of humans and VLMs across different structural patterns. [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Proportions of ’No Illusion,’ ’Human Like,’ and ’N/A’ [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 8
Figure 8. Figure 8: Effect of different prompts and fine-tuning on VLM re [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]
Figure 7
Figure 7. Figure 7: Error rates of humans and LLaVA across different con [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 11
Figure 11. Figure 11: Deception rates of LLaVA and human participants [PITH_FULL_IMAGE:figures/full_fig_p008_11.png]
Figure 10
Figure 10. Figure 10: Comparison of color term accuracy and frequency [PITH_FULL_IMAGE:figures/full_fig_p008_10.png]
Figure 14
Figure 14. Figure 14: Prompt design for generating color recognition ques [PITH_FULL_IMAGE:figures/full_fig_p011_14.png]
Figure 12
Figure 12. Figure 12: Prompt design for generating color comparison ques [PITH_FULL_IMAGE:figures/full_fig_p011_12.png]
Figure 13
Figure 13. Figure 13: Prompt design for generating color comparison ques [PITH_FULL_IMAGE:figures/full_fig_p011_13.png]
Figure 15
Figure 15. Figure 15: Website interface designed for collecting human responses to color illusion perception tasks. [PITH_FULL_IMAGE:figures/full_fig_p012_15.png]
Figure 16
Figure 16. Figure 16: Proportions of ‘No Illusion,’ ‘Human Like,’ and ‘N/A’ [PITH_FULL_IMAGE:figures/full_fig_p012_16.png]
Figure 17
Figure 17. Figure 17: Examples of image classification for color perception [PITH_FULL_IMAGE:figures/full_fig_p012_17.png]
Figure 18
Figure 18. Figure 18: This figure illustrates the proportion of responses from different models across three types of illusions on our development set. [PITH_FULL_IMAGE:figures/full_fig_p013_18.png]
Figure 19
Figure 19. Figure 19: (a) When presented with a well-known color illusion scraped from the web, GPT-4o can recognize the illusion and identify that [PITH_FULL_IMAGE:figures/full_fig_p013_19.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 36 canonical work pages

  1. [1]

    Drunk tank pink: And other unexpected forces that shape how we think, feel, and behave

    Adam Alter. Drunk tank pink: And other unexpected forces that shape how we think, feel, and behave. Penguin, 2014. 2

    work page 2014

  2. [2]

    Qwen-vl: A versatile vision-language model for un- derstanding, localization, text reading, and beyond, 2023

    Jinze Bai, Shuai Bai, Shusheng Yang, Shijie Wang, Sinan Tan, Peng Wang, Junyang Lin, Chang Zhou, and Jingren Zhou. Qwen-vl: A versatile vision-language model for un- derstanding, localization, text reading, and beyond, 2023. 5

    work page 2023

  3. [3]

    Sequential modeling enables scalable learn- ing for large vision models

    Yutong Bai, Xinyang Geng, Karttikeya Mangalam, Amir Bar, Alan L Yuille, Trevor Darrell, Jitendra Malik, and Alexei A Efros. Sequential modeling enables scalable learn- ing for large vision models. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 22861–22872, 2024. 8

    work page 2024

  4. [4]

    Understanding human perception by human-made illusions

    Claus-Christian Carbon. Understanding human perception by human-made illusions. Frontiers in human neuroscience, 8:566, 2014. 1, 2

    work page 2014

  5. [5]

    Spatialvlm: Endowing vision-language models with spatial reasoning capabilities, 2024

    Boyuan Chen, Zhuo Xu, Sean Kirmani, Brian Ichter, Danny Driess, Pete Florence, Dorsa Sadigh, Leonidas Guibas, and Fei Xia. Spatialvlm: Endowing vision-language models with spatial reasoning capabilities, 2024. 1

    work page 2024

  6. [6]

    Internvl: Scaling up vision foundation models and aligning for generic visual-linguistic tasks, 2024

    Zhe Chen, Jiannan Wu, Wenhai Wang, Weijie Su, Guo Chen, Sen Xing, Muyan Zhong, Qinglong Zhang, Xizhou Zhu, Lewei Lu, Bin Li, Ping Luo, Tong Lu, Yu Qiao, and Jifeng Dai. Internvl: Scaling up vision foundation models and aligning for generic visual-linguistic tasks, 2024. 5

    work page 2024

  7. [7]

    Reconstructing visual illusory ex- periences from human brain activity

    Fan L Cheng, Tomoyasu Horikawa, Kei Majima, Misato Tanaka, Mohamed Abdelhack, Shuntaro C Aoki, Jin Hirano, and Yukiyasu Kamitani. Reconstructing visual illusory ex- periences from human brain activity. Science Advances, 9 (46):eadj3906, 2023. 2

    work page 2023

  8. [8]

    Open visual knowledge extraction via relation-oriented mul- timodality model prompting

    Hejie Cui, Xinyu Fang, Zihan Zhang, Ran Xu, Xuan Kan, Xin Liu, Yue Yu, Manling Li, Yangqiu Song, and Carl Yang. Open visual knowledge extraction via relation-oriented mul- timodality model prompting. Advances in Neural Informa- tion Processing Systems, 36, 2024. 1

    work page 2024

Show all 42 references
  1. [9]

    Biomedical visual instruction tuning with clinician preference alignment

    Hejie Cui, Lingjun Mao, Xin Liang, Jieyu Zhang, Hui Ren, Quanzheng Li, Xiang Li, and Carl Yang. Biomedical visual instruction tuning with clinician preference alignment. Ad- vances in Neural Information Processing Systems, 2024. 9

    2024

  2. [10]

    Instructblip: Towards general- purpose vision-language models with instruction tuning,

    Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, and Steven Hoi. Instructblip: Towards general- purpose vision-language models with instruction tuning,

  3. [11]

    An image is worth 16x16 words: Transformers for image recognition at scale, 2021

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition ...

    2021

  4. [12]

    Smith, Wei-Chiu Ma, and Ranjay Krishna

    Xingyu Fu, Yushi Hu, Bangzheng Li, Yu Feng, Haoyu Wang, Xudong Lin, Dan Roth, Noah A. Smith, Wei-Chiu Ma, and Ranjay Krishna. Blink: Multimodal large language models can see but not perceive, 2024. 5, 2

    2024

  5. [13]

    Visual illusion and action

    Maurizio Gentilucci, Sergio Chieffi, Elena Daprati, M Cristina Saetti, and Ivan Toni. Visual illusion and action. Neuropsychologia, 34(5):369–376, 1996. 1, 2

    1996

  6. [14]

    Art and illusion: A study in the psychology of pictorial representation-millennium edition

    Ernst Hans Gombrich. Art and illusion: A study in the psychology of pictorial representation-millennium edition

  7. [15]

    Convolutional neural net- works can be deceived by visual illusions

    Alexander Gomez-Villa, Adrian Martin, Javier Vazquez- Corral, and Marcelo Bertalm ´ıo. Convolutional neural net- works can be deceived by visual illusions. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 12309–12317, 2019. 2

    2019

  8. [16]

    Color illusions also deceive cnns for low-level vision tasks: Analysis and implications

    Alexander Gomez-Villa, Adrian Mart ´ın, Javier Vazquez- Corral, Marcelo Bertalm´ıo, and Jes ´us Malo. Color illusions also deceive cnns for low-level vision tasks: Analysis and implications. Vision Research, 176:156–174, 2020. 2

    2020

  9. [17]

    Perceptual illusions and brain models

    Richard Langton Gregory. Perceptual illusions and brain models. Proceedings of the Royal Society of London. Series B. Biological Sciences, 171(1024):279–296, 1968. 2

    1968

  10. [18]

    Visual illusions

    Richard L Gregory. Visual illusions. Scientific American, 219(5):66–79, 1968. 1, 2

    1968

  11. [19]

    Deep residual learning for image recognition, 2015

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition, 2015. 8

    2015

  12. [20]

    Howard, Menglong Zhu, Bo Chen, Dmitry Kalenichenko, Weijun Wang, Tobias Weyand, Marco An- dreetto, and Hartwig Adam

    Andrew G. Howard, Menglong Zhu, Bo Chen, Dmitry Kalenichenko, Weijun Wang, Tobias Weyand, Marco An- dreetto, and Hartwig Adam. Mobilenets: Efficient convolu- tional neural networks for mobile vision applications, 2017. 8

    2017

  13. [21]

    Visual instruction tuning, 2023

    Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. Visual instruction tuning, 2023. 1, 5, 7

    2023

  14. [22]

    Improved baselines with visual instruction tuning, 2024

    Haotian Liu, Chunyuan Li, Yuheng Li, and Yong Jae Lee. Improved baselines with visual instruction tuning, 2024. 1

    2024

  15. [23]

    Gao, Xi Vin- cent Wang, and Lihui Wang

    Sichao Liu, Jianjing Zhang, Robert X. Gao, Xi Vin- cent Wang, and Lihui Wang. Vision-language model-driven scene understanding and robotic object manipulation. In 2024 IEEE 20th International Conference on Automation Science and Engineering (CASE), pages 21–26, 2024. 1

    2024

  16. [24]

    Art, illusion and the visual system

    Margaret S Livingstone. Art, illusion and the visual system. Scientific American, 258(1):78–85, 1988. 2

    1988

  17. [25]

    Unified-io 2: Scaling autoregressive multimodal models with vision, language, audio, and action, 2023

    Jiasen Lu, Christopher Clark, Sangho Lee, Zichen Zhang, Savya Khosla, Ryan Marten, Derek Hoiem, and Aniruddha Kembhavi. Unified-io 2: Scaling autoregressive multimodal models with vision, language, audio, and action, 2023. 5

    2023

  18. [26]

    D. L. MacAdam. Chromatic adaptation ∗. J. Opt. Soc. Am., 46(7):500–513, 1956. 8

    1956

  19. [27]

    Top-down and bottom-up neuromodulation over two different visual illusions

    O Maddaluno, A Facchin, D Zavagno, N Bolognini, R Daini, et al. Top-down and bottom-up neuromodulation over two different visual illusions. In Atti del 40th European Confer- ence on Visual Perception, pages 107–107, 2018. 2

    2018

  20. [28]

    Top-down and bottom-up: The illusory contour as a microcosm of issues in perception

    Glenn E Meyer and Susan Petry. Top-down and bottom-up: The illusory contour as a microcosm of issues in perception. In The perception of illusory contours, pages 3–20. Springer,

  21. [29]

    Cross-cultural differences in visual perception of color, illusions, depth, and pictures

    William L Phillips. Cross-cultural differences in visual perception of color, illusions, depth, and pictures. Cross- Cultural Psychology: Contemporary Themes and Perspec- tives, pages 287–308, 2019. 2

    2019

  22. [30]

    The m¨uller–lyer illusion seen by the brain: An event- related brain potentials study

    Jiang Qiu, Hong Li, Qinglin Zhang, Qiang Liu, and Fenghua Zhang. The m¨uller–lyer illusion seen by the brain: An event- related brain potentials study. Biological psychology, 77(2): 150–158, 2008. 2

    2008

  23. [31]

    Anwer, Tim Bald- win, Michael Felsberg, and Fahad S

    Hanoona Rasheed, Muhammad Maaz, Abdelrahman Shaker, Salman Khan, Hisham Cholakal, Rao M. Anwer, Tim Bald- win, Michael Felsberg, and Fahad S. Khan. Palo: A large multilingual multimodal language model. In Proceedings of the IEEE/CVF Winter Conference on Applications of Com- pu...

    2025

  24. [32]

    Brightness illusions drive a neuronal response in the primary visual cortex under top-down modulation

    Alireza Saeedi, Kun Wang, Ghazaleh Nikpourian, Andreas Bartels, Nikos K Logothetis, Nelson K Totah, and Masataka Watanabe. Brightness illusions drive a neuronal response in the primary visual cortex under top-down modulation. Na- ture Communications, 15(1):3141, 2024. 2

    2024

  25. [33]

    Illusionvqa: A challenging optical illu- sion dataset for vision language models

    Haz Sameen Shahgir, Khondker Salman Sayeed, Abhik Bhattacharjee, Wasi Uddin Ahmad, Yue Dong, and Ri- fat Shahriyar. Illusionvqa: A challenging optical illu- sion dataset for vision language models. arXiv preprint arXiv:2403.15952, 2024. 1, 2

    2024 arXiv

  26. [34]

    Very deep convo- lutional networks for large-scale image recognition, 2015

    Karen Simonyan and Andrew Zisserman. Very deep convo- lutional networks for large-scale image recognition, 2015. 8

    2015

  27. [35]

    Grounding lan- guage in multi-perspective referential communication, 2024

    Zineng Tang, Lingjun Mao, and Alane Suhr. Grounding lan- guage in multi-perspective referential communication, 2024. 1

    2024

  28. [36]

    Eyes wide shut? exploring the visual shortcomings of multimodal llms, 2024

    Shengbang Tong, Zhuang Liu, Yuexiang Zhai, Yi Ma, Yann LeCun, and Saining Xie. Eyes wide shut? exploring the visual shortcomings of multimodal llms, 2024. 5

    2024

  29. [37]

    Ofa: Unifying architectures, tasks, and modalities through a simple sequence-to-sequence learning framework, 2022

    Peng Wang, An Yang, Rui Men, Junyang Lin, Shuai Bai, Zhikang Li, Jianxin Ma, Chang Zhou, Jingren Zhou, and Hongxia Yang. Ofa: Unifying architectures, tasks, and modalities through a simple sequence-to-sequence learning framework, 2022. 5

    2022

  30. [38]

    Cogvlm: Visual expert for pretrained language models, 2024

    Weihan Wang, Qingsong Lv, Wenmeng Yu, Wenyi Hong, Ji Qi, Yan Wang, Junhui Ji, Zhuoyi Yang, Lei Zhao, Xixuan Song, Jiazheng Xu, Bin Xu, Juanzi Li, Yuxiao Dong, Ming Ding, and Jie Tang. Cogvlm: Visual expert for pretrained language models, 2024. 1, 5

    2024

  31. [39]

    Chain-of-thought prompting elicits reasoning in large language models, 2023

    Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Brian Ichter, Fei Xia, Ed Chi, Quoc Le, and Denny Zhou. Chain-of-thought prompting elicits reasoning in large language models, 2023. 1

    2023

  32. [40]

    Vision-language models for vision tasks: A survey, 2024

    Jingyi Zhang, Jiaxing Huang, Sheng Jin, and Shijian Lu. Vision-language models for vision tasks: A survey, 2024. 1

    2024

  33. [41]

    Adding conditional control to text-to-image diffusion models, 2023

    Lvmin Zhang, Anyi Rao, and Maneesh Agrawala. Adding conditional control to text-to-image diffusion models, 2023. 1

    2023

  34. [42]

    role”: “system, “content

    Yichi Zhang, Jiayi Pan, Yuchen Zhou, Rui Pan, and Joyce Chai. Grounding visual illusions in language: Do vision- language models perceive illusions like humans? In Pro- ceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 5718–5728, 2023. ...

    2023

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.