What does CLIP know about a red circle? Visual prompt engineering for VLMs
Open accessThe DOI matches its Crossref record; 14 of the 14 references it cites were found in a registry or an index.
Abstract
Large-scale Vision-Language Models, such as CLIP, learn powerful image-text representations that have found numerous applications, from zero-shot classification to text-to-image generation. Despite that, their capabilities for solving novel discriminative tasks via prompting fall behind those of large language models, such as GPT-3. Here we explore the idea of visual prompt engineering for solving computer vision tasks beyond classification by editing in image space instead of text. In particular, we discover an emergent ability of CLIP, where, by simply drawing a red circle around an object, we can direct the model’s attention to that region, while also maintaining global information. We show the power of this simple approach by achieving state-of-the-art in zero-shot referring expressions comprehension and strong performance in keypoint localization tasks. Finally, we draw attention to some potential ethical concerns of large language-vision models.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Language Models are Few-Shot Learners
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.14165 · Cited by 23,372
Learning Transferable Visual Models From Natural Language Supervision
Radford, A., Kim, J. W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., et al. (2021) · arXiv · doi:10.48550/arxiv.2103.00020 · Cited by 15,293
Learning to Prompt for Vision-Language Models
Zhou, K., Yang, J., Loy, C. C., & Liu, Z. (2022) · International Journal of Computer Vision · doi:10.1007/s11263-022-01653-1 · Cited by 3,304 · cited 1×
“Prompting VLMs is most commonly performed by prepending a set of learnable tokens to the text input [16, 21, 65, 66], vision input [19, 46, 64], or both text and vision inputs [41, 60], in order to easily steer a frozen CLIP model to solve a desired task.”
BackgroundJia, M., Tang, L., Chen, B.-C., Cardie, C., Belongie, S., Hariharan, B., & Lim, S.-N. (2022) · Lecture Notes in Computer Science · doi:10.1007/978-3-031-19827-4_41 · Cited by 1,965
Flamingo: a Visual Language Model for Few-Shot Learning
Alayrac, J.-B., Donahue, J., Luc, P., Miech, A., Barr, I., Hasson, Y., Lenc, K., Mensch, A., Millican, K., Reynolds, M., et al. (2022) · arXiv · doi:10.48550/arxiv.2204.14198 · Cited by 1,555
(2016) · doi:10.1007/978-3-319-46475-6_5 · Cited by 1,081
Li, J., Li, D., Savarese, S., & Hoi, S. (2023) · arXiv · doi:10.48550/arxiv.2301.12597 · Cited by 908 · cited 2×
“Generative VLMs like FLAMINGO [3] and BLIP [24] excel in captioning and visual question-answering tasks, but also have no way of solving pixel-level computer vision tasks.”
MethodFrom Recognition to Cognition: Visual Commonsense Reasoning
Zellers, R., Bisk, Y., Farhadi, A., & Choi, Y. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00688 · Cited by 769 · cited 1×
“Coloring regions of an image has been used for the VCR task [61], where a model is fine-tuned on annotated images [62].”
MethodReproducible Scaling Laws for Contrastive Language-Image Learning
Cherti, M., Beaumont, R., Wightman, R., Wortsman, M., Ilharco, G., Gordon, C., Schuhmann, C., Schmidt, L., & Jitsev, J. (2023) · 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52729.2023.00276 · Cited by 673 · cited 1×
“We show results for CLIP [38], OpenCLIP [10], DeCLIP [26] and SLIP [36].”
MethodTraining Verifiers to Solve Math Word Problems
Cobbe, K., Kosaraju, V., Bavarian, M., Chen, M., Jun, H., Kaiser, L., Plappert, M., Tworek, J., Hilton, J., Nakano, R., et al. (2021) · arXiv · doi:10.48550/arxiv.2110.14168 · Cited by 575 · cited 1×
“Fine-tuning LLMs can also lead to models that can generate code from docstrings [8] or solve math problems [13, 23].”
Background
Cited by
Prompt Engineering in Large Language Models
Marvin, G., Hellen, N., Jjingo, D., & Nakatumba-Nabende, J. (2024) · Algorithms for Intelligent Systems · doi:10.1007/978-981-99-7962-2_30 · Cited by 403
Foundation Models Defining a New Era in Vision: A Survey and Outlook
Awais, M., Naseer, M., Khan, S., Anwer, R. M., Cholakkal, H., Shah, M., Yang, M.-H., & Khan, F. S. (2025) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3506283 · Cited by 304
“For example, the recent work by Shtedriski et al. [214] shows the sensitivity of the CLIP model towards red circles, where simply drawing a red circle around a person’s face increases its chances of being misclassified as a murderer, suspect, or a missing person as data from new media consists of a…”
A systematic survey of prompt engineering on vision-language foundation models
Gu, J., Han, Z., Chen, S., Beirami, A., He, B., Zhang, G., Liao, R., Qin, Y., Tresp, V., & Torr, P. (2023) · arXiv · doi:10.48550/arxiv.2307.12980 · Cited by 280
ViP-LLaVA: Making Large Multimodal Models Understand Arbitrary Visual Prompts
Cai, M., Liu, H., Mustikovela, S. K., Meyer, G. P., Chai, Y., Park, D., & Lee, Y. J. (2024) · 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52733.2024.01227 · Cited by 88
“Based on our own observation and prior work [29], which shows that CLIP can understand visual markers, we directly inject the visual prompts into the original image space without any additional region-specific model designs.”
BackgroundMethodVision Language Models are blind
Rahmanzadehgervi, P., Bolton, L., Taesiri, M. R., & Nguyen, A. T. (2025) · Lecture Notes in Computer Science · doi:10.1007/978-981-96-0917-8_17 · Cited by 54
Metadata
- Authors
- Shtedritski, A., Rupprecht, C., & Vedaldi, A.
- Venue
- 2023 IEEE/CVF International Conference on Computer Vision (ICCV)
- Publisher
- IEEE
- Published
- 2023-10-01
- Pages
- 11953-11963
- OpenAlex
- W4390873481
- Cited by
- 134
- References
- 14
- FWCI
- 15.24
- Type
- conference paper
- Language
- en
- Open access
- green
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: What does CLIP know about a red circle? Visual prompt engineering for VLMs. https://app.cochat.ai/papers/10.1109/iccv51070.2023.01101
- The paper
- Shtedritski, A., Rupprecht, C., & Vedaldi, A. (2023). What does CLIP know about a red circle? Visual prompt engineering for VLMs. 2023 IEEE/CVF International Conference on Computer Vision (ICCV). https://doi.org/10.1109/iccv51070.2023.01101
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.