Explaining CLIP Zero-shot Predictions Through Concepts
Large-scale vision-language models such as CLIP have achieved remarkable success in zero-shot image recognition, yet their predictions remain largely opaque to human understanding. In contrast, Concept Bottleneck Models provide interpretable intermediate representations by reasoning through human-defined concepts, but they rely on concept supervision and lack the ability to generalize to unseen classes. We introduce EZPC that bridges these two paradigms by explaining CLIP's zero-shot predictions through human-understandable concepts. Our method projects CLIP's joint image-text embeddings into a concept space learned from language descriptions, enabling faithful and transparent explanation...
| Authors | Onat Ozdemir, Anders Christensen, Stephan Alaniz et al. |
|---|---|
| Primary category | cs.CV |
| Published | 30 March 2026 |
| Last updated at source | 30 March 2026 |
Models and datasets referencing this paper
- oonat/ezpc-embeddingsDataset
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.