How Hyperbolic Are Hyperbolic Vision-Language Models?
cs.CV, cs.AI
Submitted: 2026-11-10
Updated: 2026-11-10
Journal ref: Proceedings of the NeurIPS Workshop on Symmetry and Geometry in Neural Representations (NeurReps). PMLR, 2026
Code: https://github.com/fbyrman/hyperbolic-vlms
License: http://creativecommons.org/licenses/by/4.0/
The gist: Hyperbolic vision-language models (VLMs) have emerged as a rapidly growing area of research, reporting strong zero-shot performance alongside an interpretable embedding space in which radius orders…
Abstract
Hyperbolic vision-language models (VLMs) have emerged as a rapidly growing area of research, reporting strong zero-shot performance alongside an interpretable embedding space in which radius orders representations by generality. But how hyperbolic are hyperbolic VLMs in practice? We investigate this at the exponential map, the single operation through which curvature enters an otherwise Euclidean network. We show that across MERU, HyCoCLIP, ARGENT, and PHyCLIP, the exponential map is well approximated by the identity function, i.e., pairwise distances and angles in the pre-map Euclidean space barely change under projection, and removing the map at inference leaves evaluation scores unchanged. Motivated by this observation, we train Euclidean VLMs on CC3M from scratch under the same hierarchy losses. Over roughly the final 90% of the training steps, the training statistics are nearly indistinguishable from those of their hyperbolic counterparts, and the resulting models match the evaluation scores and recover the same radial distributions. Our findings establish Euclidean space as a sufficient substrate for hierarchical vision-language learning as it is currently practiced, and motivate new paradigms in which hyperbolic geometry can deliver on its theoretical promise. Code is available at https://github.com/fbyrman/hyperbolic-vlms.
Sources
- Microsoft COCO Captions: Data Collection and Evaluation Server
- Is the Geometry Doing the Work? An Operating-Point Audit of Hierarchy in Hyperbolic Vision-Language Models
- Fine-Grained Visual Classification of Aircraft
- Representation Learning with Contrastive Predictive Coding
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs